Wa0004.

scitation.
org/journal/apc
AIP Conference Proceedings Volume 2658
The 2nd International

Conference of Science and
Information Technology in Smart
Administration (ICSINTESA 2021)
Balikpapan, Indonesia • 20–21 October 2021

Editors • Massila Kamalrudin, Mark Robinson,
Mikio Aoyama and Richki Hardi
AIP Conference Proceedings
Volume 2658
ISBN: 978-0-7354-4290-0
ISSN: 0094-243X
scitation.org/journal/apc
The 2nd International Conference of Science
AIP Conference Proceedings and Information Technology in Smart Administration Volume 2658
(ICSINTESA 2021)
Conference collection
The 2nd International

Conference of Science and
Information Technology in Smart
Administration (ICSINTESA 2021)
Balikpapan, Indonesia Conference collection
20–21 October 2021
Editors
Massila Kamalrudin
Technical University of Malaysia Malacca, Melaka, Malaysia
Mark Robinson
The University of Texas at San Antonio, San Antonio, United States of America
Mikio Aoyama
Nanzan University, Nagoya, Japan
Richki Hardi
Universitas Mulia, Balikpapan, Indonesia
All papers have been peer reviewed.
Melville, New York, 2022

AIP Conference Proceedings Volume 2658
To learn more about AIP Conference Proceedings visit http://proceedings.aip.org
Editors
Massila Kamalrudin
Technical University of Malaysia Malacca
Melaka, 76100
Malaysia
Email: massila@utem.edu.my
Mark Robinson
The University of Texas at San Antonio
Department of Computer Science
UTSA Circle
San Antonio, TX 78249
United States of America
Email: Mark.Robinson@utsa.edu
Mikio Aoyama
Nanzan University
Department of Software Engineering
Aichi, Nagoya, Showa Ward, Yamazatocho
Nagoya
Japan
Email: mikio.aoyama@nifty.com
Richki Hardi
Universitas Mulia
Informatics
Jl. Letjend. TNI. Z.A Maulani No. 9 Damai Bahagia
Balikpapan, East Kalimantan 76114
Indonesia
Email: richki@universitasmulia.ac.id
Authorization to photocopy items for internal or personal use, beyond the free copying permitted under the 1978 U.S. Copyright Law (see
statement below), is granted by the AIP Publishing LLC for users registered with the Copyright Clearance Center (CCC) Transactional Reporting
Service, provided that the base fee of $30.00 per copy is paid directly to CCC, 222 Rosewood Drive, Danvers, MA 01923, USA: http://www.
copyright.com. For those organizations that have been granted a photocopy license by CCC, a separate system of payment has been arranged.
The fee code for users of the Transactional Reporting Services is: 978-0-7354-4290-0/22/$30.00
© 2022 AIP Publishing LLC
No claim is made to original U.S. Government works.
Permission is granted to quote from the AIP Conference Proceedings with the customary acknowledgment of the source. Republication of
an article or portions thereof (e.g., extensive excerpts, figures, tables, etc.) in original form or in translation, as well as other types of reuse (e.g.,
in course packs) require formal permission from AIP Publishing and may be subject to fees. As a courtesy, the author of the original proceedings
article should be informed of any request for republication/reuse. Permission may be obtained online using RightsLink. Locate the article online
at http://proceedings.aip.org, then simply click on the RightsLink icon/“Permissions/Reprints” link found in the article abstract. You may also
address requests to: AIP Publishing Office of Rights and Permissions, 1305 Walt Whitman Road, Suite 300, Melville, NY 11747-4300, USA;
Fax: 516-576-2450; Tel.: 516-576-2268; E-mail: rights@aip.org.
ISBN 978-0-7354-4290-0
ISSN 0094-243X
Printed in the United States of America
AIP Conference Proceedings, Volume 2658
The 2nd International Conference of Science and Information
Technology in Smart Administration (ICSINTESA 2021)
Table of Contents
Preface: 2nd International Conference of Science and Information Technology in Smart

Administration (ICSINTESA) – 2021 010001
Committees: 2nd International Conference of Science and Information Technology in Smart
Administration (ICSINTESA) – 2021 010002
PLENARY AND KEYNOTE PAPERS

SSRINDEX tool: An automated tool to measure level of software security readiness index
for remote working during Covid-19 pandemic
Massila Kamalrudin, Halimaton Hakimi, Mohd Faizal Abdollah, and Richki Hardi 020001
ARTIFICIAL INTELLIGENCE
Scenario of improving road user safety in Surabaya: A system thinking approach
Benyamin Limanto, Erma Suryani, and Rully Agus Hendrawan 030001
Enhanced chatbot security framework using MAC address authentication to improve customer
service quality
Richki Hardi, Ahmad Naim Che Pee, and Muhammad Haziq Lim Abdullah 030002
BIG DATA
Analysis of government regulations sentiment regarding the Eid al-Fitr homecoming
during Covid-19 pandemic in Indonesia using Twitter and Instagram data
Tubagus Ahmad Marzuqi, Evelline Kristiani, Indra Budi, Aris Budi Santoso, and Prabu Kresna Putra 040001
COMPUTER NETWORKS
Utilization of the Agora video broadcasting library to support remote live streaming
Lily Puspa Dewi, Agustinus Noertjahyana, Justinus Andjarwirawan, and Nico Gufron 050001
Android based remote driving for real time vehicle monitoring

Agustinus Noertjahyana, Lily Puspa Dewi, Justinus Andjarwirawan, and Nico Gufron 050002
COMPUTER SCIENCE
Comparison between Jaro-Winkler Distance Algorithm and Winnowing Algorithm in detecting
word similarities in Indonesian documents
Inte Christinawati Bu’ulolo, Melani Isabella Siregar, and Clara Fellysa Simanjuntak 060001
EDUCATIONAL TECHNOLOGY
The effectiveness of teaching technology in the era of the COVID 19 pandemic through virtual
classroom learning activities using telegram media
Gunawan, Y. Servanda, Nasruddin Bin Idris, Richki Hardi, and Jack Febrian Rusdi 070001
Improving academic creativity and community using Google Apps for education to construct
a virtual team
Sumardi, Suhartati, Nurfalah Setiawan, Richki Hardi, and Hanafi 070002
Increasing the importance of digital technology as a technopreneurship media in higher education
Richki Hardi, Agung Sakti Pribadi, Mundzir, Agustinus Noertjahyana, and Jack Febrian Rusdi 070003
The use of e-learning to increase student innovation in technopreneurship

Vidy, Richki Hardi, Yamani, and Wahyu Nur Alimyaningtias 070004
Design of Google Apps-based helpdesk ticketing system at University using extreme

programming approach
Yusuf Wibisono, Mada Aditia Wardhana, and Wisnu Hera Pamungkas 070005
INFORMATION SYSTEMS
Application of System Usability Scale (SUS) method in testing the usefulness of information
system Student Creativity Program (PKM) based on website
Silvia Oktaviani, Citra Wiguna, and Agus Priyanto 080001
Employee monitoring, productivity, privacy and harassment: Auditing technology training

and adoption in public administration
Shahid Anjum and Zahidah Ismail 080002
Improve educational marketing strategy through use of digital marketing technology

Richki Hardi, Suprijadi, Ririn Kusdyawati, and Agustinus Noertjahyana 080003
Utilization of social media application in regency SMEs performance

Ivan Armawan, Heri Pratikto, Sopiah, and Sudarmiatin 080004
An international tourist behaviour on mobile smartphone usage

Jack Febrian Rusdi, Nur Azman Abu, Sazilah Salam, Harya Gusdevi, Richki Hardi,
and Daniel Guna Nugraha 080005
MACHINE LEARNING
Comparison of support vector machine and K-nearest Neighbor algorithms in Indonesia
hoax classification
Inte Christinawati Bu’ulolo, Ivan Felix Sinaga, and Wira Epriana Ambarita 090001
Sentiment analysis of government regulations regarding the implementation

of reverse-transcriptase polymerase chain reaction (RT-PCR) during the Covid-19
pandemic in Indonesia (Case study: Air transportation mode)
A. Prisdayanti, I. Budi, A. B. Santoso, and P. K. Putra 090002
Abortion drug products classification using text mining: A case study of PT XYZ
Muhammad Faisal Mazidnianto, Anella Prisdayanti Damanik, Indra Budi, Aris Budi Santoso,
and Prabu Kresna Putra 090003
Predict the victory of the Freefire Esport Master League tournament using Naïve Bayes
Heru Pusponugroho, Arief Setyanto, and Anggit Dwi Hartanto 090004
Analysis of the application of data mining clustering model in the selection of the best study
program based on student interests at the University of Aceh Province
Yuli Asbar, Sapna Biby, Janner Simarmata, Abror, Angga Pratama, and Mutammimul Ula 090005
MATHEMATICS AND MODELLING
A modification of decomposition approach for solving non-linear quadratic differential equation;
Theory and application
Khalid Hammood AL-Jizani and Suhartati 100001
An efficient semi analytic technique for solving non-linear initial value problems
Khalid Hammood AL-Jizani 100002
A new technique of semi analytic approach α - Homo for obtaining the analytic solution to quadratic
differential equations: Theory and applications
Khalid Hammood AL-Jizani 100003
NATURAL SCIENCE
Chitosan effect in Chitosan/Hydroxyapatite/Magnesium nanocrystalline composite for bone
graft candidate
Fadhil Muhammad Tarmidzi, Marvel Shaddai Simanungkalit, and Nicole Barcelona Widodo 110001
PATTERN RECOGNITION
Design and implementation of vehicle detection system using You Only Look Once and Inverse
Perspective Mapping
Okky Putra Barus and Irvin Delano 120001
SCIENCE TECHNOLOGY FOR SOCIETY

COVID-19, SME training support by ASEAN governments and analysis of digital investments
in Brunei
Shahid Anjum and Nurul Syahidah Binti Haji Sharin 130001
SOFTWARE SYSTEMS
System analysis in virtual student assignments at University
Jamal, Riyayatsyah, Tri Sudinugraha, Richki Hardi, and Hanafi 140001
Preface: 2nd International Conference of Science and Information
Technology in Smart Administration (ICSINTESA) – 2021
Universitas Mulia, this time is the first time to hold an international ICSINTESA 2021
conference on a virtual basis; this is because the city of Balikpapan is still in a state of the
unfinished covid 19 pandemic. Universitas Mulia is still enthusiastic and trying to carry out
ICSINTESA 2021. Even though it is still in a pandemic condition, the enthusiasm and desire of
researchers to be involved in ICSINTESA Call-for-Papers is increasing.
International Conference ICSINTESA 2021 is a Call-for-Papers seminar aimed at researchers,
professionals, and academics to share research results presented in a paper. Universitas Mulia has
routinely held this activity since 2018 until now.
We are grateful for the participation of all committees, participants, editors and reviewers in
this year's ICSINTESA activity.
The International ICSINTESA 2021 conference invited speakers from the government and
academia, including the Minister of Education, Culture, Research and Technology, Nadiem Anwar
Makarim, represented by the Directorate General of Higher Education DR. Ir. Paristyanti Nur
Wardani.
The second resource person Governor of East Kalimantan DR. Isran Noor, represented by Prof.
DR. HM Aswin, M.M. Head of Bappeda of East Kalimantan Province, Mayor of Balikpapan
Rahmad Mas'ud S.E., M.E.
Next Prof. DR. Masilla Kamalruddin from the Technical University of Melaka Malaysia, DR.
Chiew Kang Leng from University Malaysia Sarawak and Prof. DR. Richardus Eko Indrajit from
Pradita University Jakarta.
The Conference proceedings span over 13 topical tracks and hosted contributions on Artificial
Intelligence, Big Data, Computer Networks, Computer Science, Digital Forensics and Security,
Educational Technology, Information Systems, Machine Learning, Mathematics and Modelling,
Natural Science, Pattern Recognition, Science Technology for Society, Software Systems, etc. On
this occasion, keynote speakers delivered expert lectures on various recent research fields like
Research challenges in Artificial Intelligence & Machine Learning and Blockchain technology.
These keynote speakers are from Indonesia and other countries.
There were 29 oral presentations by participants which brought great opportunity to share their
research findings. Efforts taken by peer reviewers contributed to improving the quality of papers
and provided constructive critical comments; improvements and corrections to the authors are
greatly appreciated.
We are thankful to all the authors who submitted papers and their presentations that really
made this conference a grand success. Last but not the least; we are thankful for the enormous
support of AIP Conference Proceedings in every step of our journey towards success.
Get along with this seminar activity, we hoped it can be an active effort of higher education to
underpin Indonesia in confronting the increasing information technology development. We wish
the technology advance is able to help human life be more effective, and efficient but still
humanist.
Best wishes
Chairman of ICSINTESA 2021
Richki Hardi
The 2nd International Conference of Science and Information Technology in Smart Administration (ICSINTESA 2021)
AIP Conf. Proc. 2658, 010001-1–010001-1; https://doi.org/10.1063/12.0012564
Published by AIP Publishing. 978-0-7354-4290-0/$30.00
010001-1
Committees
Organizing Committee
Chairperson:
Richki Hardi
Members:
Mundzir, Gunawan, Sumardi, Yustian Servanda, Yamani, Subur Anugerah, Yusuf Wibisono,
Muhammad Safi'i, Elza Rendrawati, Wisnu Hera Pamungkas, Suhartati, Erliyani, Agus
Wijayanto, Rahmat Saudi Al Fathir, Nariza Wanti Wulansari, Hasnawi, Nanda Mahardika Putra,
Nadya Ihza Maharani, Akhmad Priyanto, I Gusti Ayu Prajna Paramita, Linda Fauziyah Ariyani,
Riski Zulkarnain, Ivan Armawan, Jamal, Vidy, Dinda Annisa Prahesti.
Advisory Boards
1. Prof. Dr. Massila kamalrudin, M.Sc. (Universiti Teknikal Malaysia Melaka, Malaysia)
2. Prof. Dr. Mark Robinson (The University of Texas at San Antonio, United States)
3. Prof. Dr. Mikio Aoyama (Nanzan University Nagoya, Japan)
4. Dr. Ahmad Naim Che Pee (Universiti Teknikal Malaysia Melaka, Malaysia)
5. Dr. Khalid Al-Jizani (Mustansiriyah University, Baghdad, Iraq)
6. Dr. Agung Sakti Pribadi (Universitas Mulia, Indonesia)
7. Ir. Lukito Edi Nugroho, M.Sc., Ph.D (Universitas Gadjah Mada, Indonesia)
8. Prof. Dr. Dahlan Abdullah, M.Kom. (Universitas Malikussaleh, Indonesia)
9. Dr. Ir. Heru Purnomo Ipung, M.Eng. (Universitas Swiss German, Indonesia)
10. Jack Febrian Rusdi, S.T., M.T (Sekolah Tinggi Teknologi Bandung, Indonesia)
11. M. Zainal Arifin, S.Si., M.Kom (Universitas Negeri Malang, Indonesia)
Organizer
010002-1
SSRINDEX Tool: An Automated Tool to Measure Level of
Software Security Readiness Index for Remote Working
During Covid-19 Pandemic
Massila Kamalrudin1, a), Halimaton Hakimi1, b), Mohd Faizal Abdollah1, c), and
Richki Hardi1,2, d)
Author Affiliations
1
Faculty of Information and Communication Technology (FTMK), Universiti Teknikal Malaysia Melaka, Malaysia
2
Author Emails
a)
Corresponding author: massila@utem.edu.my
b)
halimaton.mpp1516@gmail.com
c)
faizalabdollah@utem.edu.my
d)
richki@universitasmulia.ac.id
Abstract. Resulting from the COVID-19 pandemic, which requires mankind to practice social distancing, companies and
government agencies do not have any options but to send their workers home and practice remote working. With very
limited guidance, workers are required to work remotely by utilizing various software facilitated by Internet. The increased
dependence on the cyberspace opens up the organization to become highly vulnerable to cyber threats/crimes, which may
affect their performance. Hence, organizations need to manage their cyber risk by ensuring that they are capable to manage
software security. Motivated from these problems, we have developed a prototype tool, called SSRIDEX tool (Software
Security Readiness Index tool) to measure the level of software security readiness index among workers working remotely.
This SSRINDEX tool which runs in mobile is usable to measure of the level of software security readiness in organization
especially during this COVID-19 pandemic where most workers are working remotely.
INTRODUCTION
Malaysia government need to provide effective software security infrastructure and environment for better public
sector especially during COVID-19 pandemic. Impact from the COVID-19 pandemic, which required mankind to
practice social distancing, companies and government agencies do not have options but to send their workers home
and practice remote working. The pandemic has particularly lit a fire and created an air of urgency in this “remote
migration”. Along with this active ambition there must also be an urgency to support and assist common local
communities with software security needs. The change to more prevalence and acceptance for work from home
positions is necessary and beneficial but it must not leave business processes, personal data or critical infrastructure
at risk. This is usually due to underfunding or ignoring software security all together. Working from home can end up
just as bad or worse with employees using their own unsecure network to conduct critical work processes and handle
sensitive data. The increased dependence on the cyberspace opens up the organization to become highly vulnerable to
cyber threats/crimes, which may affect organization performance. In this case, one of the ways for organizations to
manage their cyber risk is by ensuring that they are ready and aware of software security. Despite having information
020001-1
security policy in place, the readiness of organizations in ensuring the awareness among the workers and their
compliance with software security while working remotely is still questionable, thus exposing them to several security
threats. Further, this policy is too restrictive, lacking the aspect of software security. Hence, it is necessary to consider
readiness perspective to model software security public sector in order to meet requirements in the better manage
cyber threats from occurring. The identification of the readiness model enables the organizations to identify the
weakness that needs improvements; hence they can proactively manage any possible cyber threats from occurring. In
fact, the security readiness model generalized to the public sectors can help to create a more strategic cyber risk
management in future. Whereas many studies of network and technology readiness focused solely on technology
aspect, there has been a lack of model that include management and human components to build the complete picture
of an organization’s software security readiness. Further, although there are several studies and works on software
security, these studies tend to focus on addressing the security issues during the development of software and studies
that managing software securities from the end-users are still scarce. It needs to be properly managed to public
organization when they experience the security measures.
Therefore, there is a need to develop an automated tool to measure the level of security readiness among workers
working remotely. This readiness model enables the organizations to take proactive actions for continuously
improving their weaknesses related to cyber threats.
The purpose of this study to describes the tool support and usage which is able to automatically calculate and
measure the level of software security readiness index.
The rest of this paper is organized as follows: Section II presents the background and motivation. Section III
presents the conceptual framework of software security readiness model in this paper. Section IV presents tool usage
example. Section IIV concludes the paper with some discussions about automated tool for software security readiness
index and future works.
BACKGROUND AND MOTIVATION
Software Security, Cyber Risk Management and Remote Working

Software security refers to the protection of the programs that are either bought from an outsider vendor or created
in-house by users. It concerns with the methods used for controlling software used to run the operating system or
utility software (Wright, 2008). The focus of software security is to proactively protect assets from attacks that will
result in losses. Organizations that lack awareness of software security may suffer from cyberthreats which may affect
the performance of the organizations and lead to losses. Hence, operating in the vulnerable cyber environment, it is
crucial for organizations to be equipped with software security. Unlike most of the studies on software security that
focus on addressing security at the beginning of the software process, built into the design, implementing it in the
coding and verifying it during testing (Firesmith, 2012), this research aims to develop the capability of software
security among end-users.
The increasingly development of IoT devices and the existence of sophisticated attackers have resulted in the
emergence of cyber risks. Cyber risks refer to the operational risks to information and technology assets that have
consequences affecting the confidentiality, availability, or integrity of information or information systems, and it can
be classified according to the activity (e.g., criminal and non- criminal), the type of attack (e.g., malware, insider
attack, spam, distributed denial of service), and the source (e.g., terrorists, criminals, government) (Eling and Schnell,
2016). Therefore, workers working from home should have some knowledge about cyber privacy and cybersecurity,
in which failing to do so may damage the reputation of the organization.
When working remotely, employees are no longer working directly under the security of the organization.
Communication is mainly conducted via mobile applications or other channels, which increases the possible sources
of security threats (Hazari, Hargrave and Clenney, 2008). Possible issues related to the heavily dependent on the
technology are the cyber threats that may happen when working remotely are phishing scams that result in numerous
data breaches, virtual private network that could not handle everyone working from home, issues of security and data
privacy and many others (Ahmad, 2020). Hazari et al, (2008) claim that enhancing security depends on changing the
beliefs, attitudes, and behaviours of individuals and groups. According to Chan, Woon, and Kankanhalli (2005), a
breach in security does not generally result from a flaw in the technical system but it can be a result of noncompliant
employee behaviour. Based on their study of the non-compliant behaviour, they identified two main factors of non-
compliance which are individual perception climate and self-efficacy. Therefore, a study of software security readiness
among workers working from home is necessary to address the threats of cyber risks faced by organizations.
020001-2
An analysis of selected studies from 2015 indicate that there are four main security aspects related to organizational
securities which are i) information security (Kirlappos et al., 2015; Shorabi et al., 2016; Singh, 2017; Antoniou, 2018),
ii) security policy (Kirlappos et al., 2015; Shorabi et al., 2016; Sigh, 2017; Antoniou, 2018; Almubayedh et al., 2019),
iii) security training (Sari and Nurshabrina, 2016; Singh, 2017; Almubayedh et al., 2019) and iv) security awareness
(Kirlappos et al., 2015; Sari and Nurshabrina, 2016; Singh, 2017; Almubayedh et al., 2019). For the purpose of this
study, these factors will be considered as the antecedents for software security readiness.
Readiness Model/Index
Readiness referred to the state of being fully ready to engage in a specific activity. It highlights the importance of
timing, state and specificity of an activity, in relation to psychological, behavioral and structural preparedness of the
organization (Lokuge et al., 2018). In this study, readiness is considered as the state of capabilities for continuous
improvement. According to Yen et al., 2012, the measurement of readiness is crucial as it: i) provides pre- assessment
of the organization’s tangible and intangible capabilities, ii) indicates where the capability building is required and iii)
reduces the risk of failure.
To date, readiness model or indexes related to digital technology that have been developed are such as, the
Government Artificial Intelligent Index (2019) and the Global Digital Index (Cisco, 2020). The purpose of these
readiness indexes is to measure the state of capabilities of certain effort across the globe and they can be used as
benchmark for future improvement in future. Specifically, the Global Digital Index was developed to help nations
understand their position in the uptake of digitalization, to measure the progress the nation has made towards digital
maturity and demonstrate areas of strengths and provide guidance as to how they can invest to improve their overall
readiness. Another example is the Government AI readiness index, which was developed to provide an overview of
the government’s readiness to use AI. This index facilitates global comparisons as well as the ability to track
government progress in this area overtime. Focusing on examining how ready the government to implement AI based
on four generic clusters, which are governance, infrastructure and data, skills and education and innovative practices
of the government and public services.
Studies on readiness tend to relate to the use of ICT at the organizational level. There are works to related to
readiness model encompassing cloud security risk and readiness (Paulo, 2017), cybersecurity readiness index
(Kiplimo, 2018) and e-readiness assessment model (Gupta, 2015). Paulo (2017) developed a readiness model that
enables cloud services customer to tackle the needs across several different security area, such as compliance,
governance or data protection as well as allow the identification of potential risk factors related with the use of cloud
computing. However, the focus on this study is on cloud services rather than on information security. Kiplimo (2018)
proposed a web-based model that compute cyber security readiness index [CRI] for hospital based on the four pillars
of cybersecurity: people, process, policy and technology. Due to the vulnerability to attacks, there is a need to develop
a model that will help hospital perform self- assessment in determining its readiness to fight the cyber- attack menace.
Although this study has relevance to information security and cybersecurity, it is contextualized within healthcare
industry. Gupta (2015) proposed e-readiness assessment model, by defining e-readiness as the availability of the
necessary physical infrastructure, bandwidth, reliability and affordable prices, access to information [software and
information system], availability of devices required to access information [hardware and network availability and
survivability]; degree to which the ICT master plans and policies cope with the vision of an organization, the security
level adopted by the organization and the human resource to use and manage these resources for the implementation
of e-governance . This work tends to focus on the organizational level towards the implementation of e-governance,
deemphasize on readiness or awareness of software security among workers. Drawn from the selected studies above,
studies related to software security readiness index are still lacking, and this is the gap this research aims to address.
Conceptual Framework
Figure 1 shows the conceptual framework of software security readiness model for remote working in public
sector. The model is developed based on the literature review. The developments of the model start with determination
of factors of security readiness of smart governments. The purpose of developing a new software security readiness
model for remote workers as reference for top management to measure the software security readiness of organization
for level security. Based on the literature review conducted, we found that there are six factors of software security
for remote workers which are: 1) self-attitude, 2) it knowledge, 3) trust, 4) security training, 5) security policy, 6)
information security, and 7) infrastructure. The first seven (7) factors are derived from the literature review, while the
020001-3
later two factors are based on the perception of the researchers which is reliability and usability. security readiness
and they are interrelated to each other. For the individual aspect has three constructs which are self-attitude, IT
knowledge and Trust. Th organization aspect there are three constructs which are security policy, security training,
and information security. The technology aspect has three constructs which are infrastructure, reliability and usability.
Motivated from the gaps described by the literature and determine factor of software security readiness in section
2, we propose to overcome the gaps through proposing a software security readiness model for remote working in
public sector. This software security readiness model will employ the concept of readiness model design as per
described in Section 2. This is because it found that readiness model able to incorporate human perspective and
technology for better change in remote working in public organization. Figure 1 shows the conceptual framework
proposed for this study. As shown in Figure 1, the software security readiness model for remote working. It is believed
that the proposed model and index can be used as benchmark as well as comparison with other public organizations
on their software security readiness. Furthermore, the index can be used as guidance to develop policy for the
organization specifically and for the country (Cybersecurity Malaysia) for ethical use of digital technologies especially
for remote working.
Based on this conceptual model, we are developing the formulation equation to calculate and measure the level of
software security readiness index.
FIGURE 1. The Proposed Conceptual Framework Software Security Readiness Model (SOSERM)
Tool Usage Example

We have developed a prototype tool, called Software Security Readiness Index tool (SSRIDEX). The SSRIDEX
was developed using PHP programming language and adopts Model-View-Controller (MVC) design pattern and three
tier architectures. Our tool provides the (1) library for software security readiness factor, (2) Software security
readiness checklist, (3) Calculate percentage for software security readiness index and (4) Display result of level
software security readiness index. In this section, we demonstrated the features of our tool using the persona as per
described below:
Hally, a senior manager of IT departments would like to measure the readiness level of employee on software
security system in organization by using SSRIDEX tool especially employee working remotely during pandemic
Covid-19. She contacts with Salimah, who is manager and working remotely during pandemic in order to calculate
and measure their readiness on software security in organization.
As shown in Figure 2, Salimah have to sign in this system based on detailed provided by company email to ensure
the employee information was protected by company.
020001-4
FIGURE 2. Login Page
Once she successful login in the system, she can see detail information of this system and instruction step by step
on how to use this software security readiness index system. As shown in Figure 3.
FIGURE 3. Homepage system and instruction process to use this system
As shown in Figure 4, she has to read and answered the question related on software security system in her
organization/company. To answer this checklist, she needs to click the answer the checklist box. She only can click
for one answer in each question. Once she has done answer all the question, she needs to click button “Calculate
Readiness Index” to calculate and measure the readiness level.
020001-5
FIGURE 4. Software security readiness index checklist box
A shown in Figure 5, after clicking “Calculate Readiness Index” the it shows the output such as the percentage for
individual aspects, percentage for organization aspects, percentage of technology, software security readiness index
and level of software security readiness. If the employee have low or medium level of software security readiness
index. The organization have to close monitoring and give training their employee on the software security system.
Based Figure 5, the result shows that Salimah got medium level of software security readiness and she have to redo
this software checklist until she get high level of readiness.
In summary, Hally as senior manager in IT deparments use the SSRIDEX tool to realize the level of software
security readiness that automatically calculated from checklist box. Hally also uses the tool to manage and analyze
the level of software security readiness to protect assets from attacks that will result in losses. Sometimes, human error
combined with increased security threats is highlighting the need for software security awareness training. Therefore,
company must make sure every employee should go through security awareness training regularly to keep security
top of mind and reduce the risk of social engineering.
FIGURE 5. Display result and level of software security readiness
020001-6
CONCLUSION AND FUTURE WORKS
In summary, we have presented our prototype tool called as SSRIDEX that provides the (1) library for software
security readiness factor, (2) Software security readiness checklist, (3) Calculate percentage for software security
readiness index and (4) Display result of level software security readiness index. The intention to develop this tool to
help organization to calculate and measure the readiness level of their employee using software security system. This
tool may be used in continuous employee learning and development programs, which employees attend periodically
and where they are continuously evaluated by measuring readiness of software security. For future research, we will
conduct the evaluation of our tool by evaluating the efficacy of our tool in terms of usability test. The purpose of
usability test to evaluate the usefulness of our tool and to capture the feedback of the participants. We strongly believe
that software security readiness index can be used as guidance to develop proper policy for the organization
specifically and for the country (Cybersecurity Malaysia) for ethical use of digital technologies especially for remote
working. This readiness index also enables the organization to take proactive action for continuously improve their
weakness related to cyber threats.
ACKNOWLEDGMENTS
The authors would like to thank the Universiti Teknikal Malaysia Melaka for supporting this research and grant:
PASCA-COVID19/2020/FTMK-CACT/C00001.
REFERENCES
Ahmad, Tabrez, Corona Virus (COVID-19) Pandemic and Work from Home: Challenges of Cybercrimes and Cybersecurity (April
5, 2020).
Chan, M., I. Woon, and A. Kankanhalli. 2005. “Perceptions of Information Security in the Workplace: Linking Information
Security Climate to Compliant Behavior.” Journal of Information Privacy and Security 1 (3): 18–41.
Cisco. (2020). Cisco Global Digital Readiness Index 2019, White paper, available at https://www.cisco.com/c/en/us/
about/csr/research-resources/digital-readiness.html Eling, M., & Schnell, W. (2016). What do we know about cyber risk and
cyber risk insurance? The Journal of Risk Finance, 17(5), 474–491.
Fong, S., Dey, N., & Joshi, A. (Eds.). (2020). ICT Analysis and Applications. Lecture Notes in Networks and Systems.
doi:10.1007/978-981-15-0630-7
Hazari, S., Hargrave, W., & Clenney, B. (2008). An Empirical Investigation of Factors Influencing Information Security Behavior.
Journal of Information Privacy and Security, 4(4), 3–20.
Holt, D.T., Bartczak, S.E., Clark, S.W., and Trent, M.R. 2007. "The Development of an Instrument to Measure Readiness for
Knowledge Management," Knowledge Management Research & Practice (5:2), pp 75-92.
Kiplimo, A.E. (2018). A web-based model to determine cybersecurity readiness index for hospitals towards adoption of e-health,
unpublished dissertation
Kirlappos, I., Parkin, S. & Sasse, M.A., 2015. “ Shadow Security ” as a tool for the learning organization. In SIGCAS Computer
& Society, pp. 29–37.
Lokuge, S., Sedera, D., Grover, V., & Dongming, X. (2018). Organizational readiness for digital innovation: Development and
empirical calibration of a construct. Information & Management. doi:10.1016/j.im.2018.09.001
Oxford Insight (2019), Government Artificial Intelligence Readiness Index 2019, available at https://ai4d.ai/index2019/
Sun, J., Ahluwalia, P., & Koong, K. S. (2011). The more secure the better? A study of information security readiness. Industrial
Management & Data Systems, 111(4), 570– 588
Yen, H.R., Wang, W., Wei, C.-P., Hsu, S.H.-Y., and Chiu, H.-C. 2012. "Service Innovation Readiness: Dimensions and
Performance Outcome," Decision Support Systems (53:4), pp 813- 824.
020001-7
Scenario of Improving Road User Safety in Surabaya: A
System Thinking Approach
Benyamin Limanto 1, a), Erma Suryani,1, b), and Rully Agus Hendrawan,1, c)
Author Affiliations
1 Departement of
Information System, Faculty of Intelligent Electrical and Informatics Technology, Institut
Teknologi Sepuluh Nopember, Jl. Teknik Kimia, Keputih, Kec. Sukolilo, Surabaya, Jawa Timur, Indonesia
Author Emails
a)
Corresponding author: blbenyamin9@gmail.com
b)
erma.suryani@gmail.com
c)eraha@is.its.ac.id
Abstract. Road safety is one of the most crucial things in ensuring the comfort of road users. But every year, road
accidents are commonplace. Based on WHO data, road accidents occupy the eighth position as the cause of death at all
ages. The number of accidents in the city of Surabaya continues to increase every year. Based on data compiled by the
Surabaya City government, most accidents in Surabaya City are caused by driver errors. In addition, road conditions are
also one of the factors that cause accidents. The growth in the number of motorized vehicle ownership that continues to
increase every year, followed by a high accident rate, does not provide a sense of security to road users. In the city of
Surabaya itself, several efforts have been made to improve the safety of road users such as implementing the Intelligent
Transport System (ITS), training for motorized vehicle drivers, and socializing on traffic accidents, but the number of
accidents each year remains high. To overcome this problem, a systems thinking approach is used to identify the behavior
of system characteristics that affect the safety of road users. The results obtained from this study are a framework of
thinking in the form of a causal loop diagram that serves to demonstrate the interrelationships between variables and
scenarios to improve the safety of road users. The resulting Causal Loop Diagram is a model that can be used as a basis
for developing a model for improving road user safety and is expected to reduce the number of accidents, increase road
user safety, and be used as a framework for strategic planning to improve user safety.
INTRODUCTION
The safety of road users is a problem that continues to be a focus in the field of land transportation. Each year,
according to the World Health Organization (WHO), the number of deaths due to road accidents reaches 1.35
million people in the world. Deaths due to road accidents are the eighth leading cause of death for all ages. In
Southeast Asia alone, the average dea th rate percentage is 20.7 people per 100k population 1 . In Indonesia alone,
nationally since 2018 the death rate from road accidents has reached 12 people per 100k population, with a t o t a l o f
28k deaths in 2014, and is estimated to reach 65k deaths in 2035 2 . Based on this condition, the central go v ern m en t
through the transportation department prepared a national road safety plan/ RUNK in 2011 3 . However, the n u m b er
of accidents each year has continued to increase since the enactment of the RUNK from 2011 nationally with the
number of accidents reaching 2016 a total of 31262 deaths and 30694 in 2017 4 .
In the city of Surabaya itself, the number of accidents every year from 2016 to 2019 is on average above 1000
cases, with a death range between 170 to 210 fatalities and does not show the slightest downward trend, and moves
up along with road violations 5 . Factors that cause Road accidents themselves can be divided into several part s su ch
030001-1
as those caused by motorists, highways, or external consequences such as weather. In addition, the types of
accidents can be categorized into front-side collision, rear-end collision, impact with other stationary vehicles or
obstacle 6 . In the case of accidents in the city of Surabaya, almost 99% of the factors causing accidents are due to the
negligence of the drivers 5 . This condition also supported by data on the number of accidents on roads in the city of
Surabaya in 2018 the main factor causing the accident was the negligence of the driver. The drivers involved in this
accident are mostly productive ages between 15-35 years 7 . Socialization to improve road safety has been carried out
by the Surabaya City Transportation Service with the local police, but every year the number of accidents contin u es
to increase 5 .
In addition to socialization, the Surabaya City Government through the Surabaya City Smart City program, has
installed an Intelligent Transport System (ITS) with various connected technologies to support control over road
safety 8 . This extension of ITS is the emergence of an electronic ticket or e-ticket. The use of this electronic ticket
aims to reduce the number of accidents on the highway 9 . This is also supported by research which found that
controlling the number of cases of traffic violations on the smaller road will also reduce the fatality rate of acciden t s
and consecutive accidents 10 . All techniques are carried out to reduce the accident rate in the city of Surabaya both in
terms of technology and policy, but the number of accident rates continues to increase every year, so further
research is needed to identify the relationship between one variable and the variables causing accidents on the
highway to facilitate decision making in make further investments to improve the level of road safety in the city of
Surabaya.
The issue of road safety is a complex one, a nd the implementation of a single policy today can lead to
unpredictable side effects, as one variable is interrelated and reciprocals the other. The Systems Thinking technique
is an approach to solving a problem with a systems perspective that frames a re al system and is used to generate a
framework of thinking on problems dealing with complex systems that are reciprocal 11 . The output of System
Thinking is a Causal Loop Diagram model, which can be used as a causal framework for each variable in the system
and proven in the field of safety science can help way of thinking 12 . Previous research has not been able to
accommodate ITS variables, demand for transportation needs, and other factors aimed at improving road safety. Th e
result of this research is a Causal Loop Diagram to improve user safety on the highway. In the future this Causal
Loop Diagram model can be used to determine the steps that will be taken by policy makers in improving road
safety.
LITERATUR REVIEW
In the literature review section, it will be divided into 4 parts, namely the concepts of road safety and road
accidents, the concept of Systems Thinking, Causal Loop Diagrams, and previous research related to this research.
Road Safety and Road Accident
Road safety is one of the points proclaimed by the United Nations in the Sustainable Development Goals (SDGs)
that need to be achieved by every country with a global target of achieving a significant reduction of 50% 1 . In
Indonesia itself, the Indonesian National Police and the Ministry of Transportation have made the same effort by
launching the National General Plan for Road Safety (RUNK). RUNK was made based on data obtained on road
safety conditions which stated that every hour 3 to 4 people died due to road accidents 13 . Some of the Pillars
proclaimed in this RUNK are road safety management, safe roads, safe vehicles, safe road user behavior, and post -
accident handling 3 . These pillars are in line with research related to factors that cause road accidents, such as driver
factors, road factors, and external factors such as weather 6 . In addition, the increasing number of residents, the
higher the need for moving from one place to another, which ends in increasing the number of private vehicles on
road 14 , which indirectly contributes to the number of violations on road 2 .
In the city of Surabaya, this case growth average about 10-15% (from 800 at 2013 to 1348 at 2017) 15 ea ch y ea r
with accident victims are dominated by the productive population aged 15 to 35 years and dominated by young
drivers at the junior high or high school level. Most of these young riders use motorbikes as vehicles to go to school
and do not comply with driving safety standards. In addition, the unavailability of public transp ortation
infrastructure is the main cause of this happening and based on research by Widyastuti 7 the addition of a public
transportation fleet can be a solution to this problem 7 . In addition, to improve road safety, it can also be done by
issuing a ticket and increasing the number of ticket fines up to 100% from the previous fine, this has proven
effective in reducing the number of violations on the highway by 2% to 12% of the total violations, but when
030001-2
increase exceeds than 100%, this increase in fines does not create a significant reduction in accidents 16 . Drivers'
driving habits are also the key to driving safety 17 , followed by the condition of road facilities 14 , and also the use of
the Intelligent Transport System in enforcement of road safety 9 .
System Thinking
System Thinking is a way of thinking technique that uses a systems approach to solve problems in the real
world. This systems approach is carried out to look for patterns in the system that reciprocate through interactions
and the basic structure that affects the system 11 . The collection of interactions between each variable in this
complex system is difficult to predict the results, this is due to the many factors involved in the system 18 . System
Thinking forces the observer when observing the system to use an eagle eye to see the system boundaries and
communicate the system in an easy-to-understand way 19 .
According to Haines, System Thinking is simply divided into several parts, namely inputs,
throughputs/transitions, outputs, feedback loops, and the environment that affects 20 . The results of studying system s
with approaches. Using the analysis obtained, existing real systems can be formed into complex computer models 19 .
System Thinking itself has an unlimited capacity to solve problems that cannot be solved by conventional
reductionist thinking techniques, because it focuses on component relations in in the system and the components
themselves, because that relationship affects system performance 21 . To realize a safe system, System Thin k in g h a s
been proven to be able to help as a model that is used to become the basis for making decisions to reach the co n t ex t
of road safety 12 , but these models for each system will be different according to the initial conditions in the system .
To model the mental model, a Causal Loop Diagram (CLD) is used 22 .
Causal Loop Diagram
Causal Loop Diagram (CLD) is a diagram generated from a mental model that contains a series of component
variables, each of which represents a process/status 22 . Each of these variables are interrelated to form cause and
effect. This diagram is used to help understand the concepts of causality and feedback in the system, present
hypotheses about various dynamics of the system, and present the mental model of the model designer 11 . CLD
consists of variables connected by an arrow that has a polarity with a plus or minus sign which indicates a causal
effect between the variables.
+
Problem Knowledge
Articulation +
R1 +
R2 Test
+ Information
+ + B1
Question -
+ Model
FIGURE 1. Step by step to design CLD 19,22,23
The positive sign in the arrow means that the effect on the designated variable tends to increase more than it
should and if it decreases it will decrease than it should if the indicating variable decreases or increases. The
negative sign in the arrow means that the effect on the designated variable tends to decrease if the indicating variable
increases, and the designated variable will increase if the pointing varia ble decreases 23 . Final version of CLD is the
result of various iterations in formulating to answer the problem to be solved 19 . An example of a simple CLD
containing the polarity of the variables can be seen in figure 1 .
Related Works
Several previous studies have been carried out using systems thinking, but they are still lacking in depth,
especially with the emergence of the Intelligent Transportation System. Some of these studies are utilizing and
building systems thinking models to analyze the factors and causes of road accidents by Zhang et. all , and Kumar
and Umardevi 24,25 but do not pay attention to the Intelligent Transport System (ITS), research which focuses on
030001-3
road safety investment by Goh and Love 26 without considering external factors, the research focuses on traffic
safety assurance 27 but does not discuss investment factors, user driving habits, and other factors in traffic safety
assurance, research on health CLD development (humans in) transportation 28 by Harrison et. All who are trying to
improve the model Goh and Love but are still lacking in terms of training for road drivers and ITS. Based on this
search, it is necessary to develop a road safety model specifically by involving ITS, because Surabaya City has u sed
ITS 8,9 .
METHODOLOGY
The steps or methodology in this research will be based on the steps of making the Causal Loop Diagram model
from Sterman and Suryani 22,23 :
1. Problem Articulation
In the first step, the problem is developed in the system that you want to solve with the model. In this step, it will
determine the varia bles that are the key to the system that affect the problem at hand. To obtain these variables, it
can be obtained from literature studies, conducting interviews, looking at the data they have, and observing real
systems. The past data that’s looked upon are data that published by BPS and Surabaya Resort Police. We try to
connect and quantify the base problem based on those data and past resea rch 7,15.
2. Dynamic Hypothesis
In the second step a causal model is developed and illustrated with a Causal Loop Diagram (CLD). In this CLD,
all variables that have been found previously through literature review and system observations are assembled and
searched for causal relationships and are connected between variables and their polarities. In addition, each varia b le
is also categorized into existing sub-systems.
3. Structure Validation
In this step, checks are made on the existing literature that isn’t part of problem articulation, whether the
relationship between each of the existing variables is in accordance with the reality that has been found by previous
researchers. If there are variables that do not match, it is necessary to make improvements to the existing variables
so that they can represent the real conditions that exist in the system and sub -systems.
DISCUSSION
In this discussion section, it will be divided into 3 sub-sections according to the methodology that has been
discussed previously, namely articulation problems, dynamic hypotheses, and structure validation .
Problem Articulation
Road safety is a goal to be achieved by SDG 1 and also nationally by the central government with RUNK 3,13 ,
and has also become the center of attention of the Surabaya City Government 5 . All efforts have been made by the
Surabaya City Transportation and Traffic Police Office, but in recent years Lastly, the number of accidents in the
city of Surabaya continues to increase and shows a tendency to continue to increase. There is no sign of this accident
rate going down soon, while nationally there is expected to be a significant reduction in the fatality rate and road
accident rate. Based on this condition, this problem can be modeled into a causal loop diagram model.
Dynamic Hypothesis
Based on several previous studies, it can be understood that population is driving the need for transportation, this
need drives the number of private vehicle ownership in the transportation system 14 . The increase in the number of
ownerships of private vehicles indirectly affects the number of traffic violations that occur. This number of traffic
violations increases the number of fatalities on the highway and reduces the population and has an impact on the
economy, can be seen in Figure 2 loop B1. There are several causes of road accidents apart from violating traffic
rules, such as vehicle conditions and road conditions.
In addition, Surabaya City has implemented ITS as an effort to reduce the number of accidents, training for
highway drivers and increasing the number of fines also helps to reduce the number of accidents, the higher the fine,
the more deterrent to road violators, it will not be repeated. The related loops are the loops on B2, R2, R3, and B4.
These efforts to improve road safety were also influenced by investments in road safety, which are linked to the
030001-4
GDP of the City of Surabaya, and channeled into these businesses along with public transportation, and road
improvement. The related loops can be seen in Figure 2.
Population Growth
+
Population
GDP per Capita -
GDP per Capita Vehicle Inspection
Growth +
GDP + New Vehicle +
- Ownership
Number of
Fatalities +
R1 Retired Vehicle
+ B1
R4 + +
Vehicle Ownership
<Public
+ - Road Accident Transportation>
+
+ - +
Road Safety + -
Traffic Violation Traffic Congestion
+
R3 - + + -
R2 Road Factors +
B3 OverSpeeding
ITS (Intelligent + + - B2
+ + Polution
Transport) + Investment in
Traffic Safety
+ Good Road +
Alcohol and Drugs
+ + Infrastructure Traffic Penalties
Driver Training Vehicle Factors
- Bad Road
- + + Traffic Penalties
Infrastructure
Growth
+ <Traffic Penalties>
Public Age
Transportation <Vehicle
Inspection>
Protective System
FIGURE 2. Road Safety Causal Loop Diagram Solution for Surabaya
Structure Validation
Structural validation is carried out with related literature, especially in related journals, and categorizing which
variables are internal and which are external variables. Table 1 shows the endogenous and exogenous variables for
each of the sub-systems in Figure 2.
TABLE 1. Structure Validation and References
Sub Model Endogenous Variable Exogenous Variable

Road Accident Population 14 Road Factors 25
Vehicle Ownership 12 Protective System 25
Road Accident 1,12,28 Vehicle Age 25
Number of Fatalities 27 Vehicle Inspection 16
Road Safety 14,29,30
Road Investment GDP 14 , Investment 14 GDP Growth 14
ITS 31 , Driver Training5,25,29
Good Road Infrastructure 17
Bad Road Infrastructure 17
Public Transportation 7
Traffic Violation Traffic Violation 10,16 Overspeeding 25,30
Traffic Penalties 9,10 Alcohol and Drugs 25
CONCLUSION
Based on the model that has been developed, it is found that accidents in Surabaya are caused by various fact o rs
such as road conditions, driving habits, and road violations. Population needs of transportation with driving habits
030001-5
and investment take bigger par t on road accident. To overcome this problem, it is possible to implement a policy
that focuses on reducing bad habits of motorists by increasing fines and training for drivers. This model can be used
for further research in developing a simulation model using a system dynamic approach.
ACKNOWLEDGMENTS
This research was supported by the Enterprise Systems Laboratory, Department of Informat ion Systems,
Sepuluh Nopember Institute of Technology (ITS) Surabaya .
REFERENCES
1. WHO (World Health Organization), Global Status Report on Road Safety 2018, 1st ed. (WHO, 2018).
2. A. Jusuf, I.P. Nurprasetio, and A. Prihutama, J. Eng. Technol. Sci. 49, 133 (2017).
3. Republik Indonesia, Rencana Umum Nasional Keselamatan 2011-2035, 1st ed. (Departemen Perhubungan Indonesia Pusat,
Jakarta, 2010).
4. BPS, Jumlah Kecelakaan, Korban Mati, Luka Berat, Luka Ringan, Dan Kerugian Materi 2014-2016 (2019).
5. Pemerintah Kota Surabaya, Wujudkan Keselamatan Berkendara, Pemkot Bersama Polrestabes Gencar Sosialisasikan Tertib
Berlalu Lintas (2019).
6. L. Eboli, C. Forciniti, and G. Mazzulla, Transp. Res. Procedia 47, 449 (2020).
7. H. Widyastuti and A. Utami, J. Indones. Road Saf. 1, 175 (2018).
8. N. Alamsyah, T.D. Susanto, and T.C. Chou, 2016 Int. Conf. ICT Smart Soc. ICISS 2016111 (2016).
9. E.R. Nawangsari, K. Prasetya, M. Arifin, K. Khothimah, and R. Ardhana, J. Adm. Publik Public Adm. J. 10, 156 (2020).
10. R. Factor, Accid. Anal. Prev. 64, 86 (2014).
11. E. Prahasta, Systems Thinking & Pemodelan Sistem Dinamis, 1st ed. (Penerbit Informatika, Bandung, 2018).
12. R.B. Naumann, L. Sandt, W. Kumfer, S. LaJeunesse, S. Heiny, and K.H. Lich, Curr. Epidemiol. Reports 7, 343 (2020).
13. H. Widyastuti, Istiara, R.G. Thompson, R. Tay, M. Huda, F. Widiywati, and A.R. Prabawati, Procedia - Soc. Behav. Sci.
227, 530 (2016).
14. L.L. Sun, D. Liu, T. Chen, and M.T. He, Chinese J. Traumatol. - English Ed. 22, 290 (2019).
15. C. Nisa and S. Haksama, Media Penelit. dan Pengemb. Kesehat. 30, 349 (2020).
16. R. Elvik, Accid. Anal. Prev. 92, 202 (2016).
17. C. Uzondu, S. Jamson, and D. Hibberd, Transp. Res. Part F Traffic Psychol. Behav. 73, 205 (2020).
18. M. Tani, O. Papaluca, and P. Sasso, J. Open Innov. Technol. Mark. Complex. 4, (2018).
19. H.V.S. HARALDSSON and H.U. SVERDRUP, Systems Science and System Thinking in Practice: How to Develop
Qualitative and Numerical Models for Evolving Understandings of Challenges and Response (Stockholm, 2021).
20. S. Haines, Strategic and Systems Thinking: The Winning Formula, 1st ed. (System Thinking Press, Chula Vista, 2007).
21. J.P. Monat and T.F. Gannon, Am. J. Syst. Sci. 4, 11 (2015).
22. Erma Suryani, R.A. Hendrawan, and U.E. Rahmawati, Model Dan Simulasi Sistem Dinamik (Deepublish, Sleman, 2020).
23. J.D. Sterman, Business Dynamics : SYSTEMS THINKING AND MODELING FOR A COMPLEX WORLD (McGraw-Hill,
New York, 2000).
24. W.H. Zhang, H.X. Deng, and X. Bin Wang, 2010 Int. Conf. Intell. Comput. Technol. Autom. ICICTA 2010 3, 260 (2010).
25. N.. Kumar and G.Umadevi, Trasnportation Res. Board (2011).
26. Y.M. Goh and P.E.D. Love, Saf. Sci. 50, 1594 (2012).
27. V. Kolesov and A. Petrov, Transp. Res. Procedia 36, 286 (2018).
28. G. Harrison, S.M. Grant-Muller, and F.C. Hodgson, J. Transp. Heal. 22, 101138 (2021).
29. J.J. Rolison, S. Regev, S. Moutari, and A. Feeney, Accid. Anal. Prev. 115, 11 (2018).
30. Y. Shen, E. Hermans, Q. Bao, T. Brijs, and G. Wets, Accid. Anal. Prev. 138, 105484 (2020).
31. H. Yudha and T.D. Susanto, J. Phys. Conf. Ser. 1201, (2019).
030001-6
Enhanced Chatbot Security Framework Using Mac Address
Authentication to Improve Customer Service Quality
Richki Hardi1,2 a), Ahmad Naim Che Pee2, b), Muhammad Haziq Lim Abdullah2, c)
Author Affiliations
1
Human Centered Computing - Information Systems Lab (HCC-ISL) Faculty of Information and Communication
Technology (FTMK), Universiti Teknikal Malaysia Melaka, Malaysia
2
Author Emails
a)
Corresponding author: richki@universitasmulia.ac.id
b)
naim@utem.edu.my
c)
haziq@utem.edu.my
Abstract. Customer complaints are handled as part of the customer service function. The process can be used to control
customer dissatisfaction. The greater the number of customers who are pleased with the benefits of a product, the better the
customer service. Currently, chatbots are widely used in businesses to improve and simplify customer service. Its existence
is advantageous in that it allows for faster data access, allowing it to provide services to customers more quickly. Because
of this ease of access, businesses are obligated to ensure that all devices they own are secure. Based on this, the researcher
will increase the Security Framework on the Chatbot by using MAC address authentication, which can only be accessed
by the registered customer's hardware machine address, so that the chatbot menu will not be able to run if it is accessed on
a different hardware device (smartphone, laptop, or tablet). Prior to software authentication such as hybrids such as
matching name and password, unique code, and email verification, MAC address authentication on hardware is expected
to be the first stage to be authenticated by the system.
INTRODUCTION
The presence of information technology is becoming increasingly important in society; technology is used as a
tool in almost every aspect of life. Currently, chatbots are widely used in businesses to improve services. Chatbot
applications can only be used on mobile devices that are connected to the internet. Smartphones, laptops, tablets, and
other portable electronic devices are examples of mobile gadgets. Of course, in today's world, these devices are already
in use in businesses. Its existence is advantageous in that it allows for faster data access, allowing it to provide services
to customers more quickly. Because of this ease of access, businesses are obligated to ensure that all devices they own
are secure. Data security is critical in today's technological era. Even more critical than previous periods, when data
access and archiving were restricted1. With proper security, it will have an impact on the company's smooth operation
and sustainability in the short, medium, and long term. Several types of security frameworks are prepared in the
Chatbot to protect customer personal conversation data with customer service2. The security that has been
implemented is quite good, but it needs to be improved from the hardware or machine side, because security from the
software or application side appears to be able to run on a variety of hardware devices3. As a result, the researcher
will strengthen the Chatbot's security framework by employing MAC address authentication, which can only be
accessed by the registered customer's hardware machine address, so that if it is accessed on a different hardware device
(smartphone, laptop, or tablet), the chatbot menu will not be able to run. MAC address authentication on hardware is
expected to be the first stage to be authenticated by the system before
030002-1
software such as name and password matching, unique code and email verification. Several studies have attempted to
explain the chatbot system and its security, specifically: Chatbots have altered the way people think and live because
they can be present and ready to provide service assistance while performing ot her tasks at any time and from any
location4. Several previous studies have looked into how well chatbots perform. To determine the level of performance
of existing chatbots, compare them. However, it has not taken into account aspects of the user experience. One method
for deepening this aspect is to use a user-centered design approach in which the user is involved in one or more of
them. This research is inextricably linked to the use of existing research, as well as the research results5. As a guideline
for how the language should be used in chatbot design. In addition to being referred to several evaluations that must
be considered from the research results in the analysis and design of new chatbots such as ease of use, clarity,
naturalness, friendliness, robustness regarding miscommunications, and willingness to use the system again. A chatbot
is a computer program that can have intelligent conversations with users via voice or text, usually in short bursts. The
best design idea is chosen by evaluating the approach using the usability heuristic and hints from the Messenger
platform6.
The Chatbot system in banking is a computer program designed to simulate an intelligent conversation of any
banking related questions with human users via auditory or textual methods, and the smart chatbot system for this
bank will provide appropriate responses to user requests7. This is an intelligent system that will think in the same way
that humans do. This system will benefit employees by reducing their workload. Give users prompt and accurate
responses8.
LITERATUR REVIEW
Chatbots are one of the simplest and most well-known examples of intelligent Human Computer Interaction
(HCI)9. Chatbots improve significantly when productivity applications such as simple calculators, dictionaries, or even
games are embedded in them. Furthermore, the chatbot can create memos, notes, set alarms, and open programs on
the user's computer or mobile phone on demand. There were only bots before chatbots:
The invention of chatbots heralded the beginning of a new era in technology, the era of chat services. Chatbots are
virtual people who can converse with any human using textual skills in an interactive manner 10. Many cloud- based
platforms are now available for developing chatbots, including Microsoft bot framework, IBM Watson, Kore, AWS
lambda, Microsoft Azure bot service, Chatfuel, Heroku, and others. Nonetheless, all of these techniques, such as
Artificial Intelligence, Natural Language Processing (NLP), service conversion, programming, and so on, have some
drawbacks. Furthermore, banking chatbots are a key component of FinTech. Chatbots, by combining several AI
technologies, are capable of assisting or gradually replacing the jobs of banking and financial personnel. Aside from
lowering financial institutions' staff costs, it can also improve customer convenience, work efficiency, and service
quality. However, trust, data security, and personal privacy are major concerns for customers receiving financial
services11. Chatbots use AI technology, which can pose unpredictable security risks to customers if a program is
designed incorrectly, abused, or maliciously12. This paper designs and plans a Chatbot Security Control Procedure to
protect customer data security and personal privacy.
Chatbots can work in banks because they can be programmed to interact with humans, even "humanizing"
interactions to make them appear more human13. The chatbot system can also be secured in the same way that most
websites and applications are, with two-factor authentication, encryption, and so on. According to a recent review of
studies on mobile health services, architecture plays a straightforward role in data transmission 14. The system proposed
in this study is capable of filtering and processing data via a predefined health framework chatbot in order to transmit
various information to users and the health information system (HIS) and provide accurate medical services. Unlike
existing mobile health systems, chatbot-based mobile health systems not only provide general information but also
interactively provide the information needed by users, taking into account the advancement of science and
technology15. Because Internet Protocol cameras are becoming more popular and widely used, a chatbot framework
called Security Bot is required to assist users in obtaining personal detection information from cameras that can
observe 24/7 via Facebook messenger (Sbot). A system that includes a camera network, Human Detection Server
(HDS), and Sbot server is designed to build Sbot. Sbot uses the Facebook Messenger Platform to transfer information
between the user and HDS, detects humans in real time, and updates datasets captured from surveillance cameras13.
The Facebook chatbot can be used as an experimental implementation of smart home automation because it allows
users to access and control their home appliances remotely and literally from anywhere using the Raspberry Pi,
Facebook chatbot, and Google Maps API. The system describes the ease with which users can communicate and send
commands using Facebook Messenger. The benefit is that you can control your home device from any internet-
030002-2
connected device. Additional features that can be implemented for future work include machine learning, voice
commands, and Natural Language Processing (NLP). This will allow for a more intelligent approach to understanding
user input.
In recent years, there has been a rapid growth in the use of Chatbots in a variety of fields, including Health,
Marketing, Education, Support Systems, Cultural Heritage, Entertainment, and many others16. Similarly, the student-
based e learning platform, which serves as a model for managing communication and providing students with correct
answers17. Its goal is to create a system that can detect questions and use natural language processing techniques and
domain ontologies to provide appropriate answers to students via chatbots. A chatbot is an instant message account
that can provide services to users in an efficient manner by utilizing an immediate message framework. Fast Chatbot
with web and mobile applications that are less confusing and easier to install because no installation package is
required. These packages are simple to manage and disseminate18.
METHODOLOGY
Types of Research
This type of research employs qualitative and quantitative approaches, case studies, and an examination of field
conditions through the collection of data from information sources. The findings of this study can be applied in similar
situations in other locations. Case studies are intensively conducted with various data sources and are limited in time
and location li mitations. Enhance Security Framework On Chatbot Using Mac Address Authentication To Improve
Customer Service Quality is the case raised. The research period is one semester. The workflow in Figure 2 explains
four components of chatbot communication: 1) chatbot, 2) user (human), 3) input, and 4) feedback. The general public
stated that chatbots helped them save time when seeking information 19. The primary reason for using chatbots is to
make it easier to obtain information.
Flow of Investigation
Customers who will access the internet network or chatbot must pass security verification two (two) times, as
shown in Figure 1, namely verification of the device's MAC address using firewall filtering and verification of users
and passwords using the method login. The application of these various security models aims to reduce access rights
leakage in the chatbot system. If the customer only knows the user and password without registering the MAC address,
the client will be unable to access the internet network via the chatbot.
FIGURE 1. shows an activity diagram of the security framework .
The structure of the kchart research is depicted in Figure 2:
030002-3
FIGURE 2. K Chart investigation
DISCUSSION
Chatbots are becoming increasingly popular, and they are increasing business effectiveness by providing a better
experience at a lower cost20,21. Chatbots are an ecosystem that moves quickly, and new platform features have been
added over time. Hacking is becoming more useful than ever with the increasing use of information technology in all
aspects of life. Similarly, hacking is possible in chatbots; to detect collective and contextual security attacks,
Networking Chatbot must be configured as an activity. The Intelligent Chatbot service is a market leader in user
assistance and a variety of other areas. Apart from providing numerous benefits to users, this service may also pose
risks to the company. A look at the functionality and architecture of a typical chatbot service reveals the potential
dangers of chatbots. Smart chatbot services have entered the mainstream of web-based user services and will continue
to do so. Some distinct evolutionary steps, such as audio or video chatbots, are not yet commonplace, but they are
being tested. Apart from being a communication tool, this technology has numerous advantages, but it may also pose
additional IT risks to businesses that use it. Despite the fact that Chabot cannot handle all requests or provide complete
customer service. Chatbots can be attacked and must be taken seriously in terms of safety and security. The Intelligent
Chatbot service is a market leader in user assistance and a variety of other areas. Aside from providing numerous
benefits to users, this service may also pose risks to the company. A look at the functionality and architecture of a
typical chatbot service demonstrates the potential risks associated with chatbots. Smart chatbot services have entered
the mainstream of web-based user services and will continue to do so. Some distinct evolutionary steps, such as audio
or video chatbots, are not yet commonplace, but they are being tested. Apart from communication tools, this
technology has numerous advantages, but it also carries additional IT risks for businesses that use it. As the number
of cyberattacks grows, so does the knowledge and tools required to detect and analyze such attacks.
030002-4
To deal with information security incidents, an organization must have experts who can investigate security and
have a broad knowledge. The chatbot must be intelligent enough to enlist the assistance of a reputable Internet security
tool. In such cases, it is necessary to prepare the SOC (Security Operation Center) to detect incidents, intelligently
analyze evidence with appropriate security tools, and resolve it yourself or escalate it to the appropriate authorities.
Because the data transmission medium in wireless networks is broadcast radio waves, security issues require more
serious consideration. This is one of the reasons for wireless network security vulnerabilities. To secure access, abuse,
modification, and denial of services in the network and other resources, an authentication policy was implemented. A
MAC Address is a unique identifier made up of several byte numbers that is assigned to most network adapters or
Network Interface Cards (NIC). Each network device has a distinct MAC address from the others. As a result of the
MAC Address security, every network service user who wishes to connect to the network must first register his MAC
Address. This can be used to limit network service users who should not have access to it. By taking into account the
IP Address and MAC Address registered, MAC Address filtering can limit the number of computers that can connect
to the wireless hotspot. Vulnerabilities in communication line security will be more dangerous than using cables. This
necessitates the handling of additional security on a wireless network.
FIGURE 3. Interaction stages
.
FIGURE 4. Chatbot interaction process
CONCLUSION
Security Framework On Chatbot Using Mac Address Authentication To Customer Service Quality is an attempt
to improve chatbot security. The first stage of security that the system will check before moving on to the next security
stage is MAC address authentication. To interact with the chatbot, the physical address of the device that is owned by
the device must be registered. A single device has a single physical address and can only be used by one person. It is
a step toward increasing chatbot security while also educating customers about the importance of device ownership in
order for it to be maintained and maintained.
030002-5
ACKNOWLEDGMENTS
The authors would like to thank the supervisors at Universiti Teknologi Malaysia Melaka. This study was funded
by a grant from Universitas Mulia.
REFERENCES
1. I.P.G.A. Sudiatmika, KERNEL J. Ris. Inov. Bid. Inform. dan Pendidik. Inform. 1, (2021).
2. D. Huangfu, X. Tian, X. Wang, and P. Chen, ITM Web Conf. 17, (2018).
3. M. Adil, R. Khan, M.A. Almaiah, M. Al-Zahrani, M. Zakarya, M.S. Amjad, and R. Ahmed, IEEE Access 8, (2020).
4. M. Adil, M.A. Jan, S. Mastorakis, H. Song, M.M. Jadoon, S. Abbas, and A. Farouk, IEEE Internet Things J. (2021).
5. J. Korea Inst. Inf. Secur. Cryptol. 28, (2018).
6. I. Fernández-Hernández, T. Ashur, and V. Rijmen, IEEE Trans. Aerosp. Electron. Syst. 57, (2021).
7. A.C. Sari, N. Virnilia, J.T. Susanto, K.A. Phiedono, and T.K. Hartono, Adv. Sci. Technol. Eng. Syst. 5, (2020).
8. X. Song, Z. Chen, and D. Sun, J. Inf. Process. Syst. 16, (2020).
9. M. Skjuve, A. Følstad, K.I. Fostervold, and P.B. Brandtzaeg, Int. J. Hum. Comput. Stud. 149, (2021).
10. M. Chung, E. Ko, H. Joung, and S.J. Kim, J. Bus. Res. 117, (2020).
11. S. Hwang and J. Kim, Sustain. 13, (2021).
12. G. Daniel, J. Cabot, L. Deruelle, and M. Derras, IEEE Access 8, (2020).
13. S. Roca, J. Sancho, J. García, and Á. Alesanco, J. Biomed. Inform. 102, (2020).
14. M. de Gennaro, E.G. Krumhuber, and G. Lucas, Front. Psychol. 10, (2020).
15. X. Luo, S. Tong, Z. Fang, and Z. Qu, Mark. Sci. 38, (2019).
16. Z. Benyamina, K. Benahmed, and F. Bounaama, Comput. Networks 164, (2019).
17. D. Singh, B. Kumar, S. Singh, and S. Chand, Wirel. Pers. Commun. 107, (2019).
18. J. Noh, S. Jeon, and S. Cho, Electron. 9, (2020).
19. A. Ho, J. Hancock, and A.S. Miner, J. Commun. 68, (2018).
20. A.S. Pribadi, R. Hardi, Suhartati, R. Kusdyawati, and Sumardi, in J. Phys. Conf. Ser. (2021).
21. R. Hardi, in 2015 Int. Semin. Intell. Technol. Its Appl. ISITIA 2015 - Proceeding (2015).
030002-6
Analysis of Government Regulations Sentiment Regarding
the Eid Al-Fitr Homecoming During Covid-19 Pandemic in
Indonesia using Twitter and Instagram Data
Tubagus Ahmad Marzuqi1,a), Evelline Kristiani1,b), Indra Budi1,c), Aris Budi
Santoso1,d), Prabu Kresna Putra1,e)
Author Affiliation
1
Magister of Technology Information
Universitas Indonesia
Jakarta, Indonesia
Author Emails
a)
Corresponding author: tubagus.ahmad@ui.ac.id
b)
evelline.kristiani@ui.ac.id
c)
indra@cs.ui.ac.id
d)
aris.budi@ui.ac.id
e)
prabu.kresna@ui.ac.id
Abstract. The Covid-19 pandemic has resulted in an often uncertain situation. On this basis, the government has
implemented a ban on going home for the second time in 2021 to prevent a potential increase in Covid-19 cases. This rule
raises pros and cons in society. Twitter and Instagram as social media platforms then became a means to voice reactions to
the regulation, as well as opinions and criticisms The goal of this study is to find out how people feel about the situation
the "Prohibition of Homecoming in 2021". The data mining approach is used in this study to classify public sentiments
conveyed not only through the Twitter platform but also Instagram. The Naive Bayes and Decision Tree algorithms were
used to create the classification model. On Twitter data 87.93% F1 score and 92.63% F1 Score on Instagram data. This
study shows, the majority of people have negative sentiments about the "Prohibition of Homecoming in 2021" both on
Twitter and Instagram platforms.
Keywords— sentiment analysis, regulation sentiment, social media analytics
INTRODUCTION
The homecoming culture every Eid is a moment that many people have been waiting for, both Muslims and
followers of other religions. Unfortunately, the presence of the Covid-19 pandemic has made things not like they used
to be. Since the first publication of the incidence of infection in early 2020, the number of cases of the Covid-19
pandemic in Indonesia has fluctuated. Data froma shows the distribution of covid-19 cases in Indonesia. The minimum
number per day of confirmed positive cases of Covid-19 in January 2021 to April 2021 reaches 3000 cases. Ahead of
the celebration of Eid al-Fitr in 2021, the Indonesian government made a regulation regarding the ban on going home
to anticipate an increase in confirmed cases of Covid-19b.
Regulations regarding the prohibition of going home are set by the government to reap pros and cons. This rule
contradicts the custom of Indonesian Muslims who are accustomed to celebrating Eid al-Fitr in their respective
a
“Peta Sebaran COVID-19.” [Online]. Available: https://covid19.go.id/peta-sebaran-covid19. [Accessed: 16-Apr-2021].
b
D. R. Kartika, “Berlaku 6-17 Mei, Ini Aturan Lengkap Larangan Mudik Lebaran 2021.” [Online]. Available: www.kompas.com. [Accessed: 16-Apr-2021]
040001-1
hometowns. The previous year, the government issued a regulation regarding the prohibition of going home. Last
year's research 1 discussed sentiment analysis related to last year's regulations using Twitter data and found that more
people supported the regulation.
This study will conduct a sentiment analysis of the rules for the prohibition of going home this year (2021) to find
out the possible results of different sentiments from the previous year. It is possible that the results of different
sentiments can occur because of the rules regarding permits to come to tourist attractions when the homecoming ban
is enforcedc. It is hoped that the research that will be carried out can be an input for the Indonesian government in
providing rules related to going home to celebrate Eid al-Fitr during the Covid-19 pandemic in the future.
Instagram and Twitter data to be used for analysis. Instagram has the characteristics of visual-based social media,
while Twitter has the characteristics of text-based social media, this sometimes makes both of them have different
users 2. Instagram and Twitter are the top five among the most used social media platformsd. Character differences
can be a possible difference in the results of sentiment analysis on the rules. Based on that possibility we will also
compare the results between Twitter and Instagram data. So the objective of this research are:
x Differences between this year and last year sentiment analysis results of regulations regarding the prohibition
against homecoming
x Differences between the results of the sentiment analysis from Twitter and Instagram dataset
The writing in this study consists of an introduction, related works, methodology, analysis of results, conclusions,
and acknowledgments.
RELATED WORKS
Sentiment analysis is a field of study that examines a person's feelings, views, sentiments, evaluations,
judgments, attitudes, and emotions about a certain object, such as a product, service, organization, human, issue,
event, topic, or feature 3.
Currently, sentiment analysis research is widely available. Sentiment analysis 4 is to analyze the performance of
the Governor and Deputy Governor of DKI Jakarta for the 2017-2022 period based on sentiment analysis using
Twitter and Instagram data. Three algorithms were compared to choose the best algorithm, namely Naïve Bayes,
Random Forest, and SVM. The classification results shown are that the community tends to be neutral towards the
performance of the Governor and Deputy Governor of DKI Jakarta for the 2017-2022 period.
Another research is the analysis of sentiment on the policies of the Ministry of Education and Culture
(Kemendikbud) regarding the Acceptance of New Students (PPDB) who use the zoning system 5. This research uses
the K-Nearest Neighbor algorithm. It's just that this study only shows the results of the accuracy of the algorithm and
does not show the results of the sentiment analysis of these rules.
Another related research is an analysis of sentiment regarding the pros and cons of the 2019 Revision of the
Corruption Eradication Commission (KPK) Law which prompted a wave of demonstrations in September 2019 using
Twitter data 6. The algorithms used in this study are Support Vector Machine (SVM), Decision Tree, and Naïve
Bayes. Based on the sentiment analysis carried out, the results are people who speak out via Twitter and have a
negative sentiment of 75%.
Furthermore, other related research is Sentiment Analysis "Policy of Restriction of Homecoming Tradition" on
Twitter. This research was conducted during the 2020 homecoming restriction policy 1. The algorithm used is Support
Vector Machine (SVM) and Naïve Bayes. Nave Bayes produces the highest performance. And the sentiment results
also show that the majority of people support this government policy as shown by the majority of the sentiments that
have been collected are positive.
The research that will be carried out is an analysis of sentiment related to the rules for the prohibition of going
home in 2021. The results of this analysis will be compared with previous research 1, which is also about sentiment
analysis on the rules for the prohibition of going home in 2020. In addition, a comparison of sentiment analysis will
be carried out between Twitter data and Instagram data as previous study 4.
c
Henry, “Mudik Dilarang, Tempat Wisata Tetap Boleh Dibuka Selama Libur Lebaran 2021,” Liputan 6, 2021. [Online ]. Available:
https://www.liputan6.com/lifestyle/read/4524385/mudik-dilarang-tempat-wisata-tetap-boleh-dibuka-selama-libur-lebaran-2021. [Accessed: 17-Jun-2021].
d
S. KEMP, “DIGITAL 2021: INDONESIA,” 2021. [Online]. Available: https://datareportal.com/reports/digital-2021-indonesia. [Accessed: 16-Apr-2021].
040001-2
METHODOLOGY
The research takes the form of social media analytic specifically text classification. This approach was chosen with
a reason to get the sentiment of cases raised. Visually, a big picture of the stages carried out can be seen in Fig 1.
Data Collection Dataset Pre-processing Classification Model Analysis of

Twitter and Instagram Case Folding, Non-Alphanumeric Removal, Stop Naïve Bayes and Decision
with Python Word Removal, Stemming, Tokenizing Tree
Result
FIGURE 1. Research Stage

A. DATA COLLECTION
Social media data collection is done on Twitter and Instagram. Both are in the April-May 2021 period. Due to the
slightly different information display hierarchy between the two social media, the methods used are also different. To
then be processed using the Python programming language library, the Twitter dataset taken is in the form of user data,
location, and the text of tweets and retweets containing the keyword ‘dilarang mudik’ ('prohibited going home').
As for Instagram, it was decided to collect data on two specific profiles. First is @jokowi, President of the Republic
of Indonesia for this period (2019-2024) - Joko Widodo's official Instagram account. The second is @kemenhub151
which is the official Instagram account of the Ministry of Transportation of the Republic of Indonesia. Specifically
using the 'Instaloader' python library. Posts within a limited period will be screened whether they mention the issue of
homecoming restrictions. For related posts, comment data is retrieved as a data set.
B. DATA SET PRE-PROCESSING

Collected data from social media tend to be noisier than traditional print media. Mixed with the organic which is
our target, the spam, advertisements, and a variety of other unsolicited, irrelevant, or distracting material are sometimes
present. To reduce the impact of spam tweets and comments, we used basic spam filtering based on data observation
and some characteristics 4.
Furthermore, even after successfully retrieving the set of tweets and comments we want, the way of writing was
mostly informal than the meticulously edited and well-thought-out work. So it is necessary to do some pre-processing
steps 46. The preprocessing stages of this paper are:
x Case Folding, to change text into a lowercase text.
x Non-Alphanumeric Removal, to delete special characters from text, such as #, *, and other punctuation.
x Stop Word Removal, to remove stop words or words that are considered too general to have meaning for the
sentiment classification process.
x Stemming, to convert each word into the basic form of the word.
x Tokenizing, to split sentences into tokens or single words.
x Post Tagger, to Removing word types other than "ADV","JJ","ADJP","VB", "NN", "FW", and “VP”.
C. CLASSIFICATION MODEL
To make a model of classification, the dataset from the previous stage is divided into a training dataset and testing
dataset using cross-validation with k=10. The training dataset will be labeled manually by the annotator with a simple
guide for then learned by the selected algorithm models, namely Decision Tree and Naive Bayes. Naive Bayes is an
algorithm that uses Bayesian rules by applying the assumption that each attribute is conditionally independent of a
given class and generally provides competitive classification accuracy 1. In addition, naive Bayes is the best algorithm
in research on regulations regarding the prohibition of going home in 2020 1. This understanding and results make us
use the Naive Bayes algorithm.
Meanwhile, the classification of decision trees takes place in two stages, tree construction and tree pruning. To
create a tree using a top-down approach, where the tree is recursively partitioned until all data items fall into the same
label class. While tree pruning is done from the bottom up, it is done to improve the accuracy of predictions and
classification algorithms by minimizing the problem of overfitting the tree 6. This understanding makes us use the
Naive Bayes algorithm.
Feature ablation will also be used in this study. This feature is used for comparative analysis of the classification
results that have been carried out 789. First without using Post Tagger, next using Post Tagger with all accepted word
types, and finally using Post Tagger with commonly accepted word types ("JJ","VB", "NN").
040001-3
D. ANALYSIS OF RESULT
The result will be evaluated via the confusion matrix’s accuracy, precision, recall, and F-1 score. An algorithm
with the highest score and the least difference in all three output parameters is chosen. Furthermore, these results will
be analyzed to see differences in sentiment analysis last year. In addition, sentiment analysis of Twitter data and
Instagram data will be carried out based on the results that have been obtained.
ANALYSIS OF RESULT
The data obtained comes from Twitter and Instagram data. The data that has been drawn is then labeled by the
annotators. The research will refer to positive and negative labels. Positive labels are represented as supporting the
rule, while negative labels are those that do not support the rule.
Table 1 shows examples of labeling results from each annotator. The labeling process will use three annotators
(A-1, A-2 and A-3). The first annotator and the second annotator is a graduate student from Krida Wacana Christian
University. Specifically, the third annotator is a graduate student from Jenderal Sudirman University. The table also
shows examples of differences in label results between annotators 1 and 2. The third annotator will be the determinant
in determining which label will be used in the study.
TABLE 1. Example Manual Labeling

Text Data A-1 A-2 A-3 Result
@bdngfess Jangan mudik, liburan ke tempat wisata aja, ga dilarang kok Twitter + - - -
(@bdngfess Don't go home, just go on vacation to tourist attractions, it's not
forbidden)
Engga mudik pak cuma kami kangen orang tuaðŸ˜¢ðŸ˜¢ Twitter + - + +
(No going home, sir, we just miss our parentsðŸ˜¢ðŸ˜¢)
Apakah covid datang saat waktu hari raya saja? Hari hari sebelumnya covid kemana? Instagram + - - -
(Does covid only come during holidays? Where were the previous days of covid?)
Mau ke tempat pariwisata di daerah Jawa tengah aja pak.... Ga mudik ko tenang aja Instagram - + - -
pak, jangan panic
(Do you want to go to tourist places in the Central Java area, sir.... Don't go home,
don't worry, sir, don't panic)
TABLE 2. Text Preprocessing Result

Text Pre-processing Text Pre-processing Result
@korantempo Buka puasa bersama boleh. Mudik DILARANG. Selamat pagi logika yang rusak!!
Original Text (@korantempo Iftar together is allowed. Homecoming is PROHIBITED. Good morning broken
logic!!)
@korantempo buka puasa bersama boleh. mudik dilarang. selamat pagi logika yang rusak!!
Case Folding
(@korantempo iftar together is allowed. homecoming is prohibited. good morning broken logic!!)
korantempo buka puasa bersama boleh. mudik dilarang. selamat pagi logika yang rusak
Non-Alphanumeric
(korantempo iftar together is allowed. homecoming is prohibited. good morning broken logic)
korantempo buka puasa bersama boleh. mudik dilarang. selamat pagi logika rusak
Stop Word Removal
(korantempo iftar together allowed. homecoming prohibited. good morning broken logic)
korantempo buka puasa bersama boleh mudik larang. selamat pagi logika rusak
Stemming
(korantempo iftar together allow come ban good morning broken logic)
Korantempo; buka; puasa; bersama; boleh; mudik; larang; selamat; pagi; logika; rusak;
Tokenizing
(korantempo; iftar; together; allow; come; ban; good; morning; broken; logic;)
Table 2 shows an example of the results of preprocessing that has been done. The first process is Case Folding,
which is done to change uppercase letters to lowercase letters. The second process is continued with Non-alphanumeric
to remove symbols such as @, #, ! and others. Next is the Stop Word Removal process to remove stop words or words
that are considered too general to be meaningful for sentiment classification. The last two processes are stemming
which is used to remove affixes, and the tokenizing process, which is to separate sentences word by word.
Both Twitter and Instagram get higher accuracy results (i.e. the proportion of correctly classified predictions) using
the Decision Tree algorithm. Table 3 shows that the Twitter dataset processed with Naïve Bayes got 76.09% while
040001-4
the Decision Tree was 79.31%. Likewise, for the Instagram dataset, Decision Tree gets 86.47% while Naïve Bayes
gets 82.45%.
TABLE 3. Comparison of Performance

Class Twitter Instagram
Naïve Decision Naïve Decision TABLE 4. The Result of Testing Data Classification
Bayes Tree Bayes Tree Twitter Instagram
Accuracy 76,09% 79,31% 82,45% 86,47% Naïve Decision Naïve Decision
Precision + 43,28% 58,62% 31,03% 41,18 % Bayes Tree Bayes Tree
- 82,06% 80,79% 89,64% 88,16% Positive 185 57 103 31
Recall + 30,53% 17,89% 29,51% 11,48% tweets tweets comment comment
- 88,82% 96,47% 90,29% 97,57% Negative 926 1113 1129 1377
F1 Score + 35.80% 27,41% 30,25% 17,95% tweets tweets comment comment
- 85,81% 87,93% 89,96% 92,63%
Comparison of performance was measured using precision, recall, and F1 scores, as shown in Table 3. In terms of
precision show higher negative predictive values (NPV/negative class precision) on Twitter and Instagram data. On
precision, the algorithm that excels is Naïve Bayes. Recall/sensitivity shows the highest number is obtained from the
Decision Tree algorithm in the negative class of the Instagram dataset.
The F1 or F-Measure score helps to measure both Recall and Precision simultaneously, especially if there are
circumstances where it is difficult to compare two models with low precision results but high recall or vice versa. The
calculated F1-Score looks good for the negative class in both datasets. The Decision Tree algorithm has the highest
F1-score on the Instagram (92.63%) and Twitter (87.93%) datasets.
Table 4 shows that the results of all algorithms are highest on negative sentiment data on Twitter and Instagram.
This figure illustrates that the sentiments of Indonesian netizens against the rules for limiting going home are negative,
most netizens are against it.
TABLE 5. Feature Ablation

Data Feature Positive Negative Average
Precision Recall F1-Score Precision Recall F1-Score F1-pn
Twitter Post Tagger with commonly accepted
58.62% 17.89% 27.41% 80.79% 96.47% 87.94% 57.68%
word types ("JJ","VB", "NN")
Post Tagger with all accepted word types 42.86% 16.30% 23.62% 80.75% 94.17% 86.95% 55.28%
Without Post-Tagger 39.02% 17.02% 23.70% 80.20% 92.67% 85.98% 54.84%
Instagram Post Tagger with commonly accepted
41.18% 11.48% 17.95% 88.16% 97.57% 92.63% 55.29%
word types ("JJ","VB", "NN")
Post Tagger with all accepted word types 35.00% 12.07% 17.95% 88.74% 96.87% 92.63% 55.29%
Without Post-Tagger 40.00% 13.33% 20.01% 88.52% 97.09% 92.61% 56.30%
Furthermore, as shown in Table 5, ablation features were also carried out in this study. The table shows information
from the Decision Tree algorithm containing precision, recall, F1-Score of each positive and negative label, and also
shows the average F1-Score of the positive-negative label (F1-pn). In Instagram data, the F1-pn value shows 56.30%
when not using the Post-Tagger feature, then the F1-pn value drops to 55.29% when using Post-Tagger both with all
types of words and commonly accepted word types (" JJ", "VB", "NN"). Meanwhile, on Twitter data, the F1-pn value
shows 54.84% when the Post-Tagger feature is removed, then increases by 55.28% when using the Post-Tagger feature
with all types of words, when using the Post-Tagger feature with commonly accepted words types ("JJ","VB", "NN")
the F1-pn number increased by 57.68%. These results show that the use of Post-Tagger on Instagram data reduces the
value of F1-pn. But the opposite result occurs in Twitter data, the use of Post-Tagger can increase the value of F1-pn.
Although there is no opinion extraction conducted upon netizens tendency using frequently certain words - which
can be a more in-depth look at the relationship of opinion to sentiment, during the labeling process, the annotators see
the pattern of negative sentiments is related to at least three incidents that contradicted with the decision to restrict
homecoming itself.
040001-5
First, the President presence at a public figure's wedding ceremony is reinforced by the evidence contained on the
Kompas page5. Second, the government's decision to encourage the tourism sector by opening public entertainment
spaces at the same period as the decision to ban homecoming, as seen on the Liputan6 page6. Third, the 'successful
escape' of foreigners entering the country, while the Indonesian people themselves are prohibited from going home to
other regions in the country, as stated in the Kompas news page7.
CONCLUSION
Restrictions on going home during the Covid-19 pandemic in Indonesia became a decision that invited reactions
from the public. Moreover, going home to Indonesia is considered a tradition. In the previous year, with the
coronavirus which was still relatively new, last year's research conducted a focused analysis on Twitter data stating
that sentiment towards the homecoming ban was still positive. This study, which focuses on sentiment in the second
year of the imposition of the homecoming ban, turns out to find different sentiment results, the results being negative
on Twitter data and Instagram data. It is hoped that this result can be input for the Indonesian government in providing
rules regarding going home to celebrate Eid al-Fitr during the Covid-19 pandemic next year.
Sentiment analysis was carried out on Twitter data and Instagram data using the Naïve Bayes algorithm and
Decision Tree. The Decision Tree algorithm is superior to Instagram and Twitter data. The result is a negative class
of 92.63% for Instagram data and 87.93% for Twitter data.
The limitation of this research is about labeling distribution. The labeling of data between positive and negative is
not the same, with more negative labels than positive labels. This happened to both Twitter and Instagram data.
Suggestions for further research are related to the difference in text between Twitter and Instagram data because, in
this research, the ablation results show differences between Twitter and Instagram data.
REFERENCES
1
H. Suroso, I. Budi, A.B. Santoso, and P.K. Putra, 2020 3rd Int. Conf. Comput. Informatics Eng. IC2IE 2020 75
(2020).
2
B. Watkins and J.W. Lee, Commun. Brand Identity Soc. Media A Case Study Use Instagram Twitter Coll. Athl.
Brand. 476 (2018).
3
J. Xu, F. Huang, X. Zhang, S. Wang, C. Li, Z. Li, and Y. He, Appl. Soft Comput. J. 80, 387 (2019).
4
B.G. Julian, I. Budi, and D. Tanaya, ACM Int. Conf. Proceeding Ser. 122 (2019).
5
N. Anggraini and M.J. Tursina, 2019 7th Int. Conf. Cyber IT Serv. Manag. CITSM 2019 1 (2019).
6
F.S. Khurniawan and Y. Ruldeviyani, 2020 Int. Conf. Data Sci. Its Appl. ICoDSA 2020 (2020).
7
A. Kumar, S. Kohail, A. Ekbal, and C. Biemann, Lect. Notes Comput. Sci. (Including Subser. Lect. Notes Artif.
Intell. Lect. Notes Bioinformatics) 9468, 684 (2015).
8
T. Günther and L. Furrer, *SEM 2013 - 2nd Jt. Conf. Lex. Comput. Semant. 2, 328 (2013).
9
P. Chikersal, S. Poria, and E. Cambria, 647 (2015).
5
A. P. Riandi, “Polemik Kehadiran Jokowi di Pernikahan Atta Halilintar dan Aurel Hermansyah,” Kompas, 2021. [Online]. Available:
https://www.kompas.com/hype/read/2021/04/06/094736366/polemik-kehadiran-jokowi-di-pernikahan-atta-halilintar-dan-aurel-hermansyah?page=all. [Accessed: 17-
Jun-2021].
6
Henry, “Mudik Dilarang, Tempat Wisata Tetap Boleh Dibuka Selama Libur Lebaran 2021,” Liputan 6, 2021. [Online]. Available:
https://www.liputan6.com/lifestyle/read/4524385/mudik-dilarang-tempat-wisata-tetap-boleh-dibuka-selama-libur-lebaran-2021. [Accessed: 17-Jun-2021].
7
R. D. A. Saptoyo, “Ramai soal Masuknya WNA di Tengah Larangan Mudik, Ini Kata Satgas Covid-19,” Kompas, 2021. [Online]. Available:
https://www.kompas.com/tren/read/2021/05/08/160500165/ramai-soal-masuknya-wna-di-tengah-larangan-mudik-ini-kata-satgas-covid-19?page=all. [Accessed: 17-
Jun-2021].
040001-6
Utilization of the Agora Video Broadcasting Library to
Support Remote Live Streaming
Lily Puspa Dewi 1, a) Agustinus Noertjahyana 1, b) Justinus Andjarwirawan 1,c) and
Nico Gufron 1, d)
Author Affiliations
1
Petra Christian University
jl. Siwalankerto 121-131 Surabaya , Indonesia
Author Emails
a)
lily@petra.ac.id
b)
agust@petra.ac.id
c)
justin@petra.ac.id
d)
m26416094@john.petra.ac.id
Abstract. In daily life, people utilize the video technology, especially live streaming to provide a real time
transmission of an event or activity. Live streaming helps the robustness of a long-distance view by enabling the
viewer to see what would normally be seen if on-site. It describes the real time situation in video mode to fulfill the
particular information. In this case, a minimal capture-display delay is important. By using recent technological
advances have certainly made it possible to produce remote live-streaming. In this research, we utilize the Agora.io
in broadcasting live streaming video to help people view an event real time without being in the same place. We
analyzed the limitations of mobile-based live streaming applications, such as delay, frames per second, and
resolution. The paper describes experimentally the most relevant approaches for the research. The results of this
research present an overview for the public to choose the characteristics of remote live streaming.
INTRODUCTION
In daily life, people around the world have started and spent their valuable time to watch live streaming video as
a part of their daily activities. People like to communicate with others in real time through video technologies, such
as live streaming, which no longer limit their personal connection (Jyothi and Vardhan, 2016). They utilize live
streaming to provide a real time transmission of an event or activity. Live streaming helps the robustness of a long-
distance view by enabling the viewer to see what would normally be seen if on-site. It describes the real time situation
in video mode to fulfill the particular information. Live streaming refers to continuous broadcasts that are carried out
live or in real time (Lohmar, et.al, 2011).
Live streaming provides excellent opportunities for society and businesses to increase their exposure each other’s.
Live streaming also important for enterprise to build strong relationship with their customers. In today’s modern
customers, enterprise might actively share their new products, expertise and knowledge through live streaming
platforms. Enterprise can take advantage this technology to post the informative video to stay at the top of the market.
Live streaming can make major impact and make people feel different experience compared with other
communication channel such as email or an audio. The visuals can create an instant connection among people. Video
using live streaming convey the information to people more easily and efficiently.
Information reach and spread into new audience as it is appealing to cross generation with live streaming. Company
and organization recognize these trend as new emerging strategies to popularize their brand, product and services.
They can utilize live streaming as a part of marketing strategies to new customer who are not in their reach. By creating
and distribute a consistent live streaming, they can increase their new customer and your revenue exponentially. A
050001-1
particular live streaming session with customer will help understanding the information better which other
commutation channel can do like this type of interaction.
Live streaming convinces audiences to see what is happening now although they are not on site with you. The
audience feels like they are taking part in a more genuine engagement. Live streaming content can be exciting to watch
as they are not aware of what would happen in your live videos.
Based on the advantages and phenomenon as explained above, we know live streaming penetration in our daily
life is getting deeper with time. In this case, a minimal capture-display delay is important. By using recent
technological advances have certainly made it possible to produce remote live-streaming. In this research, we utilize
the Agora.io in broadcasting live streaming video to help people view an event real time without being in the same
place. We analyzed the limitations of mobile-based live streaming applications, such as delay, frames per second, and
resolution. The paper describes experimentally the most relevant approaches for the research.
LITERATURE STUDY
Live Streaming
Live streaming is an exciting technology that changes us to communicate with other people around the world. By
leveraging the incredible power of the Internet and technology, companies can easily transform their office, parlor or
any other space into global stage (Wasen, 2017). Live streaming is term that is used to describe the process of
broadcasting real time or live video footage to an audience over the Internet (Cambridge Dictionary, 2018). In
principle, live streaming is a data transferring technique where it is processed to flow as a continuous stream in real
time. Therefore, live streaming features allow a person to broadcast an event on the Internet as it occurs. Moreover,
live streaming allows one to enjoy both, audio and video, directly from the Internet without having to download it to
their computer first.
FIGURE 1. Live Streaming Workflow (Kaltura Knowledge Center, 2020)
There are five processes in live streaming workflow as present in Figure 1. The first step is capturing a good
quality video towards ensuring a smooth live streaming workflow. Some encoders have abilities to make up the
missing pieces of information in raw video, but it is better to have good quality input right at the start to smoothen up
the workflow. Good quality streaming can be achieved by optimizing the three factors of utmost importance for live
streaming which are the camera, lighting and audio. We do not need buying high end camera, all we need is a decent
camera and a well-lighted source. Some cameras have features to change the setting to get the desired output, as far
as colour coding is concerned. Finally, no less important is the audio and this needs to check microphone time and
again and reduce the background noise to make the audiences to hear source loud and clear. The next step of the
workflow is the encoder. Encoder is a must for live streaming. Encoder (can be both software or hardware) works for
compressing video and audio file in real time prior to being processed for streaming (digital format). After encoding,
the next step is converting the video into multiple display qualities for adaptive streaming, packaging the stream into
an adaptive bitrate streaming protocol, optionally encrypting the content and delivering the packaged content to the
consumption device/player typically via a CDN (Content Delivery Network). CDN helps a live streaming service to
publish audio/video faster to audiences around the globe. When the audiences try to request streaming video from a
CDN, the edge server closest to the audiences attempts to deliver it because it already has the media files cached, or
050001-2
in case it doesn’t, it will send a request to another server that does. In this way, CDNs streamline delivery of content
by streaming the videos from local servers, rather than sending them directly from the origin server each time. Hence,
a CDN uses a large network of servers placed all around the globe to distribute content quickly. The final step is
consume; the video is displayed within device (Kaltura Knowledge Center, 2020).
Agora.io
Agora is changing the way people communicate by providing audio, video and interactive broadcast SDKs that
allow to integrate interactive communications right from within your own mobile, desktop and web applications.
Agora offers capability to provide high quality communication with their core technologies i.s Real Time
Communication (RTC). RTC uses real time video and audio service as as online broadcasts. Developer might connect
to API Agora in a short time and apply video communication (Zekun, 2019).
Some of Agora.io's products are very useful for communication:
x Voice, an SDK that is useful for making voices clearer, uses 3D Spatial Audio and detects the user who is
speaking and indicates the user is talking and adds funny sound effects.
x Video, an SDK that is useful for making a video call, with internet conditions as bad as it can still make calls
clearly, able to accommodate 17 users directly and able to do customization.
x Recording, an SDK that is useful for recording in real-time with features to increase visibility, guarding
content from unwanted things.
x Live Audio Streaming, an SDK that is useful for broadcasting audio with low latency and can be listened to
across different platforms and devices.
x Live Video Streaming, an API used for integration, low latency, high quality, and broadcast audio and video.
x Agora Video Broadcasting, an SDK (Software Development Kit) used for live streaming of one-to-many and
many-to-many or live streaming video.
RESEARCH METHODOLOGY
This research analyzed the limitations live streaming applications. The paper describes experimentally the most
relevant approaches for the research. We use several specifications of smartphone as shown in Table 1.
TABLE 1. Devices
Smartphone Front Rear Memory

Operating System Display Size CPU
Device Camera Camera Internal
#1 Funtouch OS 1080 x 2340 2.0 GHz 32 MP 48 MP 128 GB,
Pixels Snapdragon 8 GB
665 Octa-core RAM
#2 Funtouch OS 1080 x 2340 1.95 GHz 25 MP 12 MP 64 GB, 6
Pixels Snapdragon GB RAM
660 AIE
Octa-core
#3 Realme UI 1080 x 2340 Qualcomm 64 MP 16 MP 128GB
RMX1922_11_C.04 Pixels Snapdragon RAM
845 2.30 Ghz 8GB
The system design of this research is using the system architecture as presented in Figure 2.
050001-3
FIGURE 2. System Architecture
System architecture in this project can be explained as follows:

1. Set the role of the client as host or audience. Host publish the video to the channel, and the audiences join the
channel.
2. Get a token from server. Joining the channel needs a token as credential that authenticates a host or audiences.
3. Join a channel.
4. Publish and subscribe. Host can publish video after join a channel.
RESULT AND DISCCUSSION
This part reflects on the findings of how live streaming performance related to the limitation of live streaming
applications.
The testing scenarios are conducted using mp4 (bitrate 256kbps) which is captured by webcam with network
background traffic 50Mbps, 75Mbps and 100Mbps. Table 2 shows the average delay on various devices for testing.
TABLE 2. Average delay
Device Resolution Bit per Average Delay (ms) Average Delay (ms) Average Delay (ms)
# second Background traffic : 50 Background traffic : Background traffic :
Mbps 75 Mbps 100 Mbps
1 640 x 360 (360p) 30 131.153 134.323 156.231
720 x 480 (480p) 30 153.232 159.344 180.224
1920 x 1080 (1080p) 30 182.123 184.211 175.835
2 640 x 360 (360p) 30 175.245 172.567 195.347
720 x 480 (480p) 30 190.345 194.298 237.345
1920 x 1080 (1080p) 30 243.619 236.235 135.137
3 640 x 360 (360p) 30 132.913 134.987 138.328
720 x 480 (480p) 30 132.949 135.751 182.345
1920 x 1080 (1080p) 30 178.321 184.987 190.761
050001-4
CONCLUSION
Real time live streaming requires a clear and smooth internet connection. Network background traffic has
important influence in this situation. Congestion points on the internet are a serious barrier to real time performance
live streaming. This bottleneck point is what makes the delay so that the audio/video is random or blurry. Hardware
specifications, especially memory, are also an influence on the smoothness of live streaming. Using agora provide
reliable live streaming video.
REFERENCES
1. Cambridge Dictionary, Meaning of “livestream” in the English Dictionary (Cambridge University Press, 2021).
Accessible: https://dictionary.cambridge.org/dictionary/english/livestream (01 February 2021)
2. Kaltura Knowledge Center, Kaltura live streaming overview. Accessible: https://knowledge.kaltura.com/kaltura-
live-streaming-overview (06 March 2021)
3. J. Wasen, Live streaming your business: A guide for the creative entrepreneur 1st edition (Gigee Inc: Missoula,
Montana. 2017).
4. S. N. Jyothi and K. V. Vardhan, "Design and implementation of real time security surveillance system using
IoT," in 2016 International Conference on Communication and Electronics Systems (ICCES). (IEEE, 2016), pp.
1-5, doi: 10.1109/CESYS.2016.7890003.
5. T. Lohmar, T. Einarsson, P. Fröjdh, F. Gabin and M. Kampmann, "Dynamic adaptive HTTP streaming of live
content," in 2011 IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks.
(IEEE, 2011), pp. 1-8, doi: 10.1109/WoWMoM.2011.5986186.
6. W. Zekun, “How Agora.io Became the Leading Player in Video and Voice,” Retrieved from
https://equalocean.com/ai/20190412-dancing (2019, 12 April).
7. Z. Lu, H. Xia, S. Heo, and D. Wigdor, “You watch, you give, and you engage: A study of live streaming practices
in China,” in CHI 2018, Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems.
(ACM, April 2018), pp. 1-13. https://doi.org/10.1145/3173574.3174040
050001-5
Android Based Remote Driving for Real Time Vehicle
Monitoring
Agustinus Noertjahyana1, a) Lily Puspa Dewi 1, b) Justinus Andjarwirawan 1, c) and

Nico Gufron 1)
Author Affiliations
1
Petra Christian University
Jl. Siwalankerto 121-131 Surabaya , Indonesia
Author Emails
a)
Corresponding author: agust@petra.ac.id
b)
lily@petra.ac.id
c)
justin@petra.ac.id
Abstract. At this time, where internet connections are increasingly being used to support daily life, they are also
widely used to monitor events. It is inseparable from every incident that occurs in transportation facilities, including
motorbikes or cars. To be able to monitor every incident that occurs when using a vehicle, this study applies an
Android-based DVR application that is equipped with GPS, speedo meter and Live Streaming. It is hoped that with
this application, users can record videos and take photos accompanied by the location and description of each incident.
In addition, the live streaming feature can help other users to join by entering the channel name. The tests carried out
show that the live streaming feature can be implemented properly using each channel.
INTRODUCTION
Vehicles have now become an important need for humans to carry out daily activities. Many functions of the
vehicle in supporting human activities such as transportation of goods and traveling from one place to another. It is
also easier for the public to get a Driving License (SIM) with online services. This of course also has an impact on the
increasing number of vehicle drivers passing on the highway.
Problems arise when more and more novice drivers often ignore road rules, such as speed limits and traffic signs.
Based on data taken from the Korlantas Polri, it is stated that the number of accidents is increasing day by day due to
the carelessness of drivers, especially exceeding the speed limit. This is common among teenagers who have just
received a Driving License.
For this reason, by utilizing Digital Video Recorder technology, many parents monitor the speed of vehicles driven
by teenagers while on the highway. However, there are limitations when using DVR devices, the difficulty of
conducting real-time monitoring due to limited access to DVRs that are currently widely sold in the market.
There are many benefits of DVR when applied in many ways, one example is the car DVR, which is an onboard
camera that can record continuously, by installing it on the dashboard of the vehicle or on the rear windshield of the
vehicle. The main use of a car DVR is that it can be used as evidence in the event of an accident and the driver is
innocent. The image recording process will take place continuously while the vehicle is being driven, and the results
of video recording depend on the quality of the DVR camera used and the recording time depends on the memory
capacity of the DVR device.
The limitation that most complained about was the difficulty in supervising the speed of the vehicle from the
supervised driver, in this case, it could be a child or a driver on goods transport services. For this reason, considering
050002-1
the sophistication of current cell phone cameras, which are generally of good quality and with GPS and data access
features, the idea of using a cell phone as a DVR is one of the most suitable alternatives to overcome the limitations
of a car DVR.
In this study, we will try to use a cellphone to be used as a DVR as well as a remote live streaming, so it is hoped
that it can replace the role of a car DVR in conducting vehicle surveillance, as well as speed monitoring by utilizing
the GPS feature of the cellphone.
Applications developed using Android-based mobile phones are currently widely used by people from young to
old, so it is hoped that this application can bring more benefits.
With the live streaming feature and being able to record along the way, and can be accessed by more than 1 person,
it is possible for parents to supervise teenagers and transportation business owners can also carry out direct supervision
of the driver.
THEORY
DVR is an electronic device capable of recording video in digital format and stored in disk drives, USB flash
drives, SD memory cards, and other storage areas. DVR is also known as PVR (Personal Video Recorder). (Engelbert
et al., 2011). The recorder is connected to a media input device, such as a satellite or cable box and television, this
unit is a large storage unit with file management software and a remote control for recording an event. (Apriyani et
al., 2020)
GPS is a satellite-based navigation owned by the United States government and operated by the United States Air
Force. GPS is ubiquitous in society in various forms of application. GPS is a hardware and software technology that
continues to become more accessible and affordable for consumers, with multiple GPS receivers. The use of GPS in
forestry has been hampered by the challenges of receiving and maintaining adequate satellite reception due to canopy
and landform interference. (Han, 2016)
Speedometer or speed meter is a measurement tool that measures and displays the instantaneous speed of a vehicle.
Now deployed in all motor vehicles, speedometers became available as an option in 1900, and as standard equipment
from 1910 onwards. (Han, 2016)
Speed is an important piece of information that is necessary for the driver in the right way. In addition to visual,
auditory, and vestibular cues, the speedometer on the dashboard provides accurate speed readings to the driver, and
provides assistance to the driver, so that the driver can accurately assess the speed at which they are traveling. (Wang
et al., 2015)
Streaming is a technology for playing video or audio files directly or with a recorder from a server machine
(webserver). In other words, video files that are on the server machine or server computer can be directly executed
and played by the client while the request process is done so that the client does not need to spend so much time
downloading the video and audio files. (Noertjahyana et al., 2019)
To apply video streaming to the network, it is necessary first to calculate the available bandwidth, to support data
transmission. Bandwidth is an important parameter for streaming on a network. The greater the available bandwidth,
the better the quality of the displayed video. (Jumisko-Pyykkö et al., 2013)
Agora.io is a global real-time service provider for video and audio. The main technology is a real-time
communication or Real Time Communication (RTC), which uses real-time voice and video services and online
broadcasts to create a scenario like offline communication. (Wisnu & Barja, 2021)
SYSTEM DESIGN
The DVR application developed in this study aims to record along the journey and can track the driver through the
speedometer. So, the driver can be reminded when driving at a speed more than normal. All events that occur while
the vehicle is running will be recorded according to the existing memory storage capacity of the device. With the help
of the Google Maps feature, the application can track the driver's location in real time and the driver can use Google
Maps to get to the destination quickly and well-directed.
Drivers can use the app to record trips, and capture important events on the way, and with live streaming, other
people who have the right and can log into the system will be able to see the driver's camera. For example, the driver's
parents can supervise their child who is driving a vehicle. There is a live streaming feature using a smartphone, so
when the driver starts driving the vehicle can press the start button which aims to start recording. The recording quality
050002-2
depends on the smartphone used for recording, as well as the length of the recording time depending on the available
smartphone memory capacity. Video files can be saved in date and time format. Results Photos and Videos will be
saved in different folders, to make it easier to find photos and videos. This application allows the device to be remote
through selecting the same channel. So that parents can monitor their children through the channels that have been
set. The channel selection process at the beginning of the application can be seen in Figure 1.
Fig. 1. Channel Selection Diagram
The next stage is the main stage of the application, namely the process of selecting a role activity. In the process
of selecting this role activity there are 3 activities, namely: Broadcaster, Audience and DVR. Broadcaster if the user
wants to record through the application, while Audience if the user wants to see the broadcaster's display. In the DVR
option, the user can record every event that occurs in the storage memory. This process can be seen in Figure 2.
050002-3
Fig 2. Role Activity Diagram
IMPLEMENTATION
The implementation of the application will be tested on an Android-based smartphone. The type of smartphone
that will be used as a device for testing the application is the Samsung S7 Edge. The Start page of the application from
seen in Figure 3.
Fig 3. Main Program

When the user starts the application, there will be a choice of Broadcaster, Audience and DVR. If the user selects
Broadcaster, then the user can do a live stream, just by entering the channel name that matches what they want. This
050002-4
feature will take place when the user is logged in on the Live Streaming page. Users with broadcaster roles have the
right to switch front or rear cameras, mute the sound but keep the live stream running. The quality of the live stream
depends on the user's internet speed. If the Broadcaster chooses a stream resolution of 1280 x 720, then the audience
will try to adjust to the same resolution at the same speed and bitrate. This live streaming test is shown in Figure 4.
Testing the live streaming feature is divided into 3 categories: Good, Moderate and Poor. Good, if the video quality is
smooth and not broken. Moderate, if the video quality is smooth but a little bit delay. Poor, if the video quality is not
good. The test results can be seen in Table 1.
Fig 4. Broadcaster Role
FPS Publish Live Streaming

View Live Streaming
Delay (Frames Bitrate (Upload speed) /
Resolution (Download speed) / Information
(seconds) Per (kbps) Broadcaster
Audience
Second)
320 x 240 1 15 200 60 kbps – 90 kbps 75 kbps – 85 kbps Moderate
480 x 360 1 30 980 130 kbps – 160 kbps 135 kbps – 150 kbps Moderate
640 x 360 1 30 1.200 160 kbps – 180 kbps 160 kbps – 185 kbps Moderate
640 x 480 1 30 1.500 190 kbps – 210 kbps 190 kbps – 245 kbps Moderate
960 x 720 2 30 2.760 300 kbps – 415 kbps 250 kbps – 280 kbps Good
1280 x 720 2 30 3.420 350 kbps – 500 kbps 250 kbps – 270 kbps Good
CONCLUSION
From the discussion that has been carried out in this study, conclusions can be drawn:
x The application can display videos well at a high-resolution of 1280 x 720.
x By utilizing agora.io, the application can be accessed together by utilizing the channel.
REFERENCES
Apriyani, S., Subagio, R. T., & Ilham, W. (2020). Perancangan Aplikasi Monitoring Ruangan Menggunakan IP
Camera Berbasis Android. Jurnal SISKOM-KB (Sistem Komputer Dan Kecerdasan Buatan), 4(1), 1–7.
https://doi.org/10.47970/siskom-kb.v4i1.161
Engelbert, B., Blanken, M., Kruthoff-Brüwer, R., & Morisse, K. (2011). A user supporting personal video recorder
based on a generic Bayesian classifier and social network recommendations. Communications in Computer
and Information Science, 185 CCIS(PART 2), 1–8. https://doi.org/10.1007/978-3-642-22309-9_1
Han, I. (2016). Car speed estimation based on cross-ratio using video data of car-mounted camera (black box).
Forensic Science International, 269, 89–96. https://doi.org/10.1016/j.forsciint.2016.11.014
Jumisko-Pyykkö, S., Haustola, T., Boev, A., Gotchev, A., Kawashima, K., Okamoto, J., Ishikawa, K., Negishi, K.,
Bouten, N., Famaey, J., Latre, S., Huysegems, R., Vleeschauwer, B. D., Leekwijck, W. V., Turck, F. D.,
Hewage, C. T. E. R., Martini, M. G., Dumi, E., Grgi, S., … Blvd, W. C. (2013). A Dynamic Adaptive HTTP
Streaming Video Service for Google Android A Dynamic Adaptive HTTP Streaming Video Service for
Google Android. Signal Processing: Image Communication, 27(4).
Noertjahyana, A., Surjo, G. M., & Palit, H. N. (2019). Streaming Media Implementation in Moodle-Based E-
Leaming Application. TIMES-ICON 2018 - 3rd Technology Innovation Management and Engineering Science
International Conference. https://doi.org/10.1109/TIMES-iCON.2018.8621669
050002-5
Wang, D., Pei, M., & Zhu, L. (2015). Detecting driver use of mobile phone based on in-car camera. Proceedings -
2014 10th International Conference on Computational Intelligence and Security, CIS 2014, 148–151.
https://doi.org/10.1109/CIS.2014.12
Wisnu, P., & Barja, M. (2021). The Students ’ Perspectives on Implementing the Mobile-based Video Calls Using
AgoraIO in Face-to-face Distance Learning Activities. 15(3), 429–437.
050002-6
Comparison Between Jaro-Winkler Distance Algorithm
and Winnowing Algorithm in Detecting Word Similarities
in Indonesian Documents
Inte Christinawati Bu’ulolo1,a), Melani Isabella Siregar1, Clara Fellysa

Simanjuntak1
Author Affiliations
1
Del Institute of Technology, Jl. PIDel, Sitoluama, Laguboti, Toba Regency, North Sumatra,
Indonesia
Author Emails
a) Corresponding author: inte@del.ac.id
Abstract. Plagiarism is copying one's work and making it seem as if it is his own work, so it can be said to be a crime
that is very rampant in the academic world in Indonesia. The scope in plagiarism is very wide, one of which is the
similarity of the contents of a document containing words, sentences or paragraphs taken from the document or writings
of others. The initial level in minimizing the act of plagiarizing in a document, namely by detecting the similarity of the
document by performing calculations using algorithms that can detect the similarity of words in the document. Therefore,
to solve the problem, it is necessary to detect the similarity of words in the document. Researchers performed several
stages starting from the preprocessing stage, then carried out the process of calculating similarities using Jaro Winkler
distance algorithm and winnowing algorithm. The calculation process is done manually and using the simulator. After
testing the two algorithms, the Winnowing Algorithm is better than the Jaro Winkler Distance Algorithm seen from the
accuracy value obtained by the Winnowing Algorithm of 94.4% and for the Jaro Winkler Distance Algorithm of 88.8%.
INTRODUCTION
Plagiarism or plagiarism is someone's work and makes it look as if the work itself is a crime that is very rampant
in the academic world in Indonesia. The act of plagiarism is getting easier to do, supported by technological
developments such as the internet which is growing and increasingly sophisticated, so it is very easily accessible by
anyone and is very helpful in working on coursework, practicum reports, journals, final assignments and others.
However, it is easy to have a negative impact on students, namely plagiarism. The negative impact of plagiarism can
result in a student's lack of integrity, one of which is bringing up new ideas and making students lazy to think.
Things that are classified as plagiarism are making as if the ideas, ideas, and works of others are the result of their
own work, taking several articles without making a reference from where the writing was taken. The scope of
plagiarism is very broad, one of which is the similarity of the contents of a document containing words, sentences or
paragraphs taken from other people's documents or writings.
Plagiarism based on percentage is divided into 3, namely [1]:
x Mild plagiarism: the similarity rate is below 30%
x Moderate plagiarism: the similarity rate is between 30-70%
x Heavy or total plagiarism: the similarity rate is above 70%
Based on a brief explanation of the problems above which are the reasons and basis for this research by detecting
similarities in Indonesian language documents with the data used obtained from journal documents, to determine the
amount of similarity in the documents being tested so that it can be concluded that the documents are the same or
not based on the percentage value. similarity.
In this study, to check the similarity of words in two documents, researchers used the Jaro Winkler Distance
algorithm and the Winnowing algorithm. The Jaro Winkler distance algorithm is a variation of the Jaro distance
metric which measures the similarity between two strings using a string metric approach. The process in the Jaro
Winkler DistanceTheAlgorithm
2nd International Conference of Science and Information Technology in Smart Administration (ICSINTESA 2021)
isAIP
calculating the length of the string, determining the number of existing characters,
Conf. Proc. 2658, 060001-1–060001-6; https://doi.org/10.1063/5.0111181
060001-1
then calculating the transposition value, calculating the Jaro Winkler Distance value to calculate the highest level of
similarity [2].
Winnowing algorithm can detect plagiarism with document fingerprinting method. The process for obtaining
fingerprinting documents uses hashing techniques. The hashing technique is carried out by the Rolling Hash process
by changing the gram string series obtained in the n-gram process into hash values. Next, it will enter the window
formation process to form the window value from a series of hash values. The minimum window value obtained in
this process is called the document fingerprint [3].
The object of this research is the Jaro Winkler Algorithm and Winnowing Algorithm. For the comparison
between the Jaro Winkler Distance Algorithm and the Winnowing Algorithm, it is not yet known the best results
from the comparison of the two algorithms. Therefore, in this study the Jaro Winkler Distance Algorithm and
Winnowing Algorithm were chosen to get the best algorithm. The data used for this study were obtained from
journal documents. The data is taken from the website https://garuda.ristekbrin.go.id/. The scope compared to the
research in this test is the abstract part of the journal document. The purpose of this research is to get the best
algorithm in detecting the similarity of words in Indonesian language documents by comparing the Jaro Winkler
Distance Algorithm with the Winnowing Algorithm.
In previous studies, the research conducted by ladder in conducting comparative analysis of Lavensthein
Distance and Jaro Winkler algorithms for Text Document Plagiarism Detection Applications. Jaro-Winkler got the
highest similarity value, which was 80.92% [2]. In the study of the use of N-Gram and Jaro-Winkler Distance in
Online Class Applications for Plagiarism Detection, it was found that the Jaro-Winkler Distance algorithm only
tested the similarity of words and could not identify the same meaning in words [4]. In the study of the application
of the Winnowing Algorithm to Detect Text Similarities in the Student Final Project, it was stated that the
Winnowing algorithm, the higher the value of n-gram, the lower the value of similarity, and vice versa, if the value
of n-gram is low, the value of similarity is higher [5]. In the Comparison of Document Plagiarism Detection Results
with the Jaro Winkler Distance Method and the Latent Semantic Analysis Method, the Jaro-Winkler algorithm has
the largest percentage, which is 100% [6]. And in the Comparison study of the Winnowing Algorithm with the
Rabin Karp Algorithm for Detecting Plagiarism in the Similarity of Thesis Title Text, the algorithm has a higher
percentage of 88.89% [3].
METHODOLOGY
The methodology used in this study to obtain the best algorithm between the Jaro Winkler Distance algorithm
and the Winnowing algorithm consist of data collection, text preprocessing, word similarity detection using Jaro
Winkler Distance Algorithm and Winnowing Algorithm and get the best Algorithm.
A. Data Collection
In this study, researchers used 3 categories of data. Each data in each category consists of 2 documents. The two
documents in each data will be compared to get the percentage similarity value. The categories used are:
1. Category All Words are not Equal

In this category all the words in the document are not the same. in this category using 1 data.
The naming of the data used is DU-1.
2. Category Partial Words Same
In this category, some of the words in the document are the same, the documents being compared are
documents that have the same object of research in the journal. To find out if the same word exists, it is
checked manually. In this category, there are variations in the data to be tested. The data variations in this
category are the displacement of the sentence position, the change in the sentence pattern from the active
sentence to the passive sentence, he displacement of the word position, the effect of the same initial word
and the influence of the match distance range. This category uses 11 data.
3. Category All Words Same
In this category, the contents of the document use the same journal, so the same words and the number of
words in document 1 and document 2 are the same. The variations used in this category are the
060001-2
displacement of the sentence position, the displacement of the word position, the effect of the beginning of
the same word and the influence of the match distance range. This category uses 6 data.
B. Text Preprocessing
The next stage after data collection is complete, the data will be preprocessed.
1. Jaro Winkler Distance Algorithm

Text preprocessing performed on the Jaro Winkler Distance algorithm, namely the Case Folding process is
a process that converts capital letters in documents into lowercase letters (a-z), omitted characters other
than letters, such as numbers and punctuation marks. Tokenizing is the process of cutting text into chunks
of words known as tokens. The stage to remove words that have low information from the text. Stopwords
are common words that usually appear in large numbers and are considered meaningless. Stemming, is the
process of finding the root word of each word by removing all affixes [3].
2. Winnowing Algorithm
Text preprocessing carried out on the Winnowing algorithm is the preprocessing stage used, namely, case
folding is a process that converts capital letters in documents into lowercase letters (az), characters other
than letters are removed, such as numbers and punctuation marks., stopwords removal/filtering is the stage
for remove words that have low information from the text. Stopwords are common words that usually
appear in large numbers and are considered meaningless., stemming is the process of finding the root word
of each word by removing all affixes, and, whitespace insensitivity is the process of removing all spaces in
the document [7].
C. Word Similarity Detection

After preprocessing the data, the two documents in each data category in each category will be compared
using the Jaro Winkler Distance algorithm and the Winnowing algorithm to get the percentage similarity value.
1. Jaro Winkler Distance Algorithm

The workings of the Jaro Winkler Distance algorithm is to calculate the jaro similarity value in the two
documents that have gone through the preprocessing process, calculate the jaro winkler distance value and
calculate the percentage. To calculate the Jaro similarity value, there are several variable values used,
namely the number of words in each of the two documents, the same number of words in both documents,
and the number of transposition values in both documents [10]. To calculate the jaro winkler distance
value, the variable values used are the jaro similarity value, the prefix value (the same number of words at
the beginning of the document before the inequality was found, the maximum value = 4) and the constant
value (p = 0.1). after getting the jaro winkler distance the results are multiplied by 100% to get the final
percentage result [2].
2. Winnowing Algorithm
The way the Winnowing algorithm works is to change the contents of the two documents into a series of n-
grams, then change the words in the n-gram sequence to a hash value using the Rolling hash method, then
the hash value will be divided into several windows with a size of W. window is the process of forming a
substring from the hash value along w-grams and from the winnowing process will produce a fingerprint
(minimum value of window formation) then the Jaccard Coefficient is carried out to calculate the word
equation after the fingerprinting process [7].
D. Get the Best Algorithm

After getting the percentage value of word similarity in each data using the Jaro Winkler algorithm and the
Winnowing algorithm, the next step is to find the best algorithm between the Jaro Winkler Distance algorithm and
the Winnowing algorithm. The steps taken to get the best algorithm are calculating accuracy and calculating time
complexity.
060001-3
x To find out the best algorithm, an accuracy calculation is carried out [8] using confusion matrix. Accuracy
is the comparison of correctly identified cases with the total number of cases [9]. Data labelling used to
calculate accuracy is data with light, medium and high plagiarism. Benchmarking used for accuracy
calculations is by manually counting the same word pairs in both documents. To calculate the percentage of
the same word manually using the formula :
Formula 1

ൌ ‫ͲͲͳכʹכ‬Ψ
ͳ൅ ʹ
Formula 1. Count the same word pairs manually on two documents
x Calculating the time complexity of the Jaro Winkler Distance algorithm and the Winnowing algorithm
RESULT AND DISCUSSION

In the results and discussion chapter, it contains the results and calculations of accuracy and time complexity on
the Jaro Winkler Distance algorithm and the Winnowing Algorithm.
A. The results and discussion chapters will show the results of the accuracy of word similarity detection from
the Jaro Winkler Distance algorithm and the Winnowing algorithm. MS Word Feature percentage results
on each data obtained from the calculation results of the Formula 1.
1. Test results and calculation of the Jaro Winkler Distance Algorithm
TABLE 1. Test results with the Jaro winkler distance algorithm and MS Word Feature Results
Jaro Winkler
MS Word Feature Distance
Data Label Label
Results Algorithm
Results
DU-1 0% R 0% R
DU-2a 37.15% S 46.2% S
DU-2b 33.51% S 40.8% R
DU-2c 37.15% S 43.4% S
DU-2d 37.15% S 45.2% S
DU-2e 37.15% S 40.8%% R
DU-2f 37.15% S 51.6% S
DU-2g 1.04% R 34% S
DU-2h 1.04% R 0% R
DU-2i 44.24% S 50.8% S
DU-2j 54.86% S 54.2% S
DU-2k 54.86% S 54.1% S
DU-3a 100% B 100% B
DU-3b 100% B 70% S
DU-3c 100% B 94.8% B
DU-3d 100% B 100% B
DU-3e 100% B 98% B
DU-3f 100% B 96% B
Information:
R = Wispy
060001-4
S = Medium
B = Heavy
So that the results:
TR =2
TS = 9
TB = 5
FRB = 0
FRS = 1
FBR = 0
FBS = 1
FSR = 0
FSB = 0
Then the accuracy result,
ൌͳ͸ȀͳͺൈͳͲͲΨൌͺͺǤͺΨ
2. Test results and calculation of the Winnowing Algorithm
TABLE 2. Test results with Winnowing algorithm and MS Word Feature Results
Data MS Word Feature Results Label Winnowing Algorithm Results Label

DU-1 0% R 0% R
DU-2a 37.15% S 31.16% S
DU-2b 33.51% S 31.62% S
DU-2c 37.15% S 28.57% R
DU-2d 37.15% S 30.48% S
DU-2e 37.15% S 30.82% S
DU-2f 37.15% S 30.72% S
DU-2g 1.04% R 20.71% R
DU-2h 1.04% R 21.15% R
DU-2i 44.24% S 31.21% S
DU-2j 54.86% S 48.47% S
DU-2k 54.86% S 47.97% S
DU-3a 100% B 100% B
DU-3b 100% B 97.79% B
DU-3c 100% B 92.35% B
DU-3d 100% B 100% B
DU-3e 100% B 98.9% B
DU-3f 100% B 96.69% B
Information:
R = Wispy
S = Medium
B = Heavy
So that the results:
TR =3
TS = 8
TB = 6
FRB = 0
FRS = 0
FSR = 1
060001-5
FBR = 0
FBS = 0
FSB = 0
Then the accuracy result,
ൌͳ͹ȀͳͺൈͳͲͲΨൌͻͶǤͶΨ
From the explanation of the accuracy values for the Jaro Winkler Distance Algorithm and Winnowing Algorithm,
the accuracy results for the Jaro Winkler Distance Algorithm have an average value of 88.8% and for the
Winnowing Algorithm it has an average of 94.4%.
B. Time Complexity
The time complexity obtained by the Jaro Winkler Distance algorithm is O(n2) and the time complexity
obtained from the Winnowing algorithm is O(n2).
CONCLUSION
Based on the results of the analysis and discussion carried out, the conclusions of this research are as follows.
1. The accuracy results obtained from the Jaro Winkler Distance Algorithm are 88.8% and the Winnowing
Algorithm is 94.4%, so the best algorithm in detecting the similarity of words in documents is the
Winnowing Algorithm.
2. Jaro Winkler Distance Algorithm and Winnowing Algorithm have the same time complexity, namely O(n2).
ACKNOWLEDGMENTS
The author is very grateful to Research and Community Services Unit (LPPM) Institut Teknologi Del for supporting
this publication.
REFERENCES
1. Sudigdo, S., 2007. Beberapa catatan tentang plagiarisme. s.l.:Majalah Kedokteran Indonesia.
2. Tannga, M. J., S. R. & . H., 2017. Analisis Perbandingan Algoritma Levenshthein Distance dan Jaro Winkler
untuk Aplikasi Deteksi Plagiarisme Dokumen Teks. JTRISTE, Volume 4, p. 44~54.
3. Sunardi, Yudhana, A. & Mukaromah, L. A., 2017. Perancangan aplikasi deteksi plagiarisme karya ilmiah
menggunakan algoritma winnowing. ISSN, Volume 1.
4. Hakim, L., 2019. Penggunaan N-Gram dan Jaro Winkler Distance pada aplikasi kelas daring untuk deteksi
plagiat. e-ISSN.
5. Wibowo, R. K. & Hastuti, K., 2016. Penerapan Algoritma Winnowing untuk Mendeteksi Kemiripan Teks pada
Tugas Akhirr Mahasiswa. Techno.Com, Volume 15.
6. T. & Elizabeth, T., 2018. Perbandingan Hasil Deteksi Plagiarisme Dokumen dengan Metode Jaro Winkler
Distance dan Metode Latent Semantic Analysis. Jurnal Teknologi dan Sistem Komputer, pp. 7-12.
7. Fauzi, R. M. & Wibawa, J. C., n.d. Implementasi Algoritma Winnowing untuk Mendeteksi Kemiripan Teks
pada Artikel. UKI.
8. Sameen, S. et al., 2017. Measuring Short Text Reuse For the Urdu Language. IEEE Access.
9. Arini, Wardhhani, L. K. & Octaviano, D., 2020. Perbandingan Aeleksi Fitur Term Frequency & Tri-Gram
Character Menggunakan Algoritma Naive Bayes Classifier (Nbc) Pada Tweet. P-ISSN, Volume 9.
10. Winkler, W. E., 1990. String Comparator Metric and Enchanced Decision Ules in the Fellegi-Sunter Model of
ecord Linkage. Proceedings of the Section on Survey Research Methods. American Statistical Association, pp.
354-359.
060001-6
The Effectiveness of Teaching Technology in the Era of the
COVID 19 Pandemic Through Virtual Classroom Learning
Activities Using Telegram Media
Gunawan1,a), Y Servanda1,b), Nasruddin Bin Idris1,c), Richki Hardi1, d), Jack Febrian
Rusdi2,e)
Author Affiliations
1
Department of Informatics, Universitas Mulia, Balikpapan, Indonesia
2
Department of Informatics, Sekolah Tinggi Teknologi Bandung, Bandung, Indonesia.
Author Emails
a)
gunawan@universitasmulia.ac.id
b)
yustians@gmail.com
c)
nasruddin@universitasmulia.ac.id
d)
e)
jack@sttbandung.ac.id
Abstract. The COVID-19 pandemic has changed various aspects of human life today, especially in education. It requires
all elements of education to adapt and continue learning in the following semester. The application of technology in
education improves the ability of teachers to deliver material to students, makes it easier for students to understand the
material presented and makes time more effective and efficient. Technology has a significant role in the field of education.
The purpose of this study is to provide choices about the effectiveness of learning using telegram, learning technology
media devices during the COVID-19 pandemic. This study uses a qualitative method, international articles, and similar
sources related to learning solutions during a pandemic using telegram media. Online learning using telegram media is an
effective alternative to activate classes even though schools have been closed considering that time and place are at risk
during the current pandemic. However, it is essential to evaluate this learning technique according to local conditions
considering that the distribution of facilities and the ability of parents to provide different online learning facilities to
students in Indonesia is still minimal.
INTRODUCTION
To prevent the Covid-19 pandemic, the government issued a policy to ask their students to study at home. Starting
March 16, 2020, schools will apply online student learning methods, then how effective is the learning. In this study,
several learning technologies that have been used in the Covid 19 pandemic era will be investigated, as well as virtual
classroom learning activities using telegram media. Currently, Corona is a hot topic of discussion. In any hemisphere,
Corona still dominates the public sphere. In a short time, his name became a trending topic, discussed here and there,
and was reported massively in print and electronic media. Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-
COV-2), better known as the corona virus, is a coronavirus of coronavirus that causes infectious disease in humans.
Covid-19 is an infectious disease caused by a newly discovered type of coronavirus. Although it primarily attacks
the elderly, this virus can strike anyone, from infants, children, to adults. This coronavirus can cause mild disorders
of the respiratory system, severe lung infections, and even death. Corona Virus Disease was first discovered in Wuhan,
070001-1
China, in December 2019. This virus spreads very quickly and has spread to almost all countries, including Indonesia,
in just a few months. So that on March 11, 2020, WHO declared this outbreak a global pandemic. It has made several
countries set policies to impose lockdowns to prevent the spread of the coronavirus. In Indonesia itself, the Large-
Scale Social Restriction approach was implemented to suppress the reach of this virus. Because Indonesia is currently
implementing PSBB, all activities carried out outside the home must be stopped until this pandemic subsides.
Some local governments have decided to implement policies to leave students and start implementing online
learning methods (online) or online. This government policy came into effect in several provinces in Indonesia on
Monday, March 16, 2020, which was also followed by other areas. But this does not apply to some schools in each
region. These schools are not ready with an online learning system, requiring learning media such as mobile phones,
laptops, or computers1. The online learning system (on the network) is a learning system without face to face directly
between teachers and students but carried out online using the internet network2. Teachers must ensure that teaching
and learning activities continue, even though students are home3. The solution, teachers are required to design learning
media as an innovation by utilizing online media (online)4.
It follows the Minister of Education and Culture of the Republic of Indonesia regarding Circular Letter Number 4
of 2020 concerning the Implementation of Educational Policies in the Emergency Period for the Spread of Corona
Virus Disease (COVID-19). This study can implement the learning system through a personal computer (PC) or laptop
connected to an internet network connection. Teachers can do learning together at the same time using groups and
chat and virtual meetings using telegrams like other media as learning media5. Thus, the teacher can ensure that
students take part in learning simultaneously, even though in different places6.
LITERATUR REVIEW
In the literature review section, the discussion will be divided into three sections: the concept of learning
communication media, online learning, and the effectiveness of digital-based learning based on previous research
studies.
Learning Communication Media

Communication media are all means used to produce, reproduce, distribute or disseminate and convey information,
for that communication media plays a significant role in people's lives. The process of sending data in modern times
is very sophisticated, so telecommunications technology is most sought after to convey or transmit information or
news because telecommunications technology is growing, getting faster, more precise, accurate, easy, cheap, effective
and efficient. 7.
Meanwhile, communication media in education and learning include human-based media, print-based media,
visual-based media, Audio-Visual-based media, computer-based media, and the use of libraries as learning resources.
In this case, communication has an essential role in teaching so that communication between educators and students
goes well and students can accept the information conveyed by educators, and educators need to use learning media8.
The existence of communication plays an essential role in teaching. considering that communication between
teachers and students is going well and students can accept the information conveyed by the teacher, teachers need to
use learning media, and if you see the benefits of communication media in education and learning, including; help the
learning process to achieve learning objectives, motivate students, present information with youth, stimulate
discussion, direct student activities, carry out exercises and tests, strengthen learning, provide simulation experiences9.
Communication media provides more useful value so that education has a positive use-value for life and life 10.
Communication as a science, where the structure of science or knowledge itself includes epistemological,
ontological, axiological aspects 11. Meanwhile, if reviewed, that educational communication can be interpreted as
communication that occurs in an academic atmosphere 12. Thus, scholarly communication involves travelling
messages or information that penetrates the field or educational events. Communication is no longer accessible but
controlled and conditioned for educational purposes in this sphere13.
In education, communication is a crucial determinant in achieving goals13. Because no matter how clever and broad
an educator's knowledge is, if he cannot communicate his thoughts, knowledge, and insights well, of course, he will
not be able to transform his knowledge to his students. In comparison, communication media is a medium that is
useful as an intermediary that delivers information from the source and recipient14. Examples of communication media
include television, film, radio, photos, audio recordings, projected images, printed materials. In this case, educational
media is one of the influential supporters in helping the learning process occur, so that in the learning process, teaching
070001-2
media is a container and channel for messages from the source of the news, in this case, the teacher, to the recipient
of the message, the student15. Limitations of teaching media as everything that can use to stimulate students' thoughts,
feelings, attention, and willingness to encourage the learning process in students16.
By seeing how important the position of communication media in education is, the demands to follow up on
knowledge in terms of its benefits are very many, including through research, articles, papers, papers. Whatever this
paper is as an aspect of research, it only focuses on the benefits of communication media in education and the teaching
knowledge of an educator. If they cannot communicate their thoughts, expertise and insights well, they will not be
able to transform their knowledge to their students17. While the communication media is a medium that is useful as an
intermediary that delivers information from the source and recipient. Examples of communication media include
television, film, radio, photos, audio recordings, projected images, printed materials. In this case, educational media
is one of the influential supporters in helping the learning process occur, so that in the learning process, teaching media
is a container and channel for messages from the source of the news, in this case, the teacher, to the recipient of the
message, the student18.
In a broader sense, the definition of teaching media is anything that can stimulate students' thoughts, feelings,
attention, and willingness to encourage the learning process in students19. By seeing how important the position of
communication media in education is, the demands to follow up on knowledge in terms of its benefits are numerous,
including through research, articles, papers, papers20. Whatever this paper is as one aspect of the study, it only focuses
on the benefits of communication media in education and teaching21. One of the communication media that is often
used in learning is online media, such as what's app, email, classroom, and telegram. 22. The presence of telegram
makes a significant change because it makes it easier for netizens to carry out various activities by taking advantage
of the advantages possessed by telegram23, as well as changing one's perspective to become global in addressing
matters related to learning, especially if, as a digital citizen, you must be able to take advantage of this change
appropriately. Useful and responsible24.
For this reason, the use of telegrams as a medium of communication in learning is one solution during conditions
that require face-to-face learning to be replaced with online learning25.
Telegram
Telegram has long been popular before the smartphone era. Telegram was a post office facility used to send long-
distance written messages quickly. But after technology developed rapidly, this facility was eroded and not used
anymore. Now Telegram's name was taken by a startup developed into an app. Telegram is a cloud-based instant
messaging application that focuses on speed and security. Telegram is designed to make it easier for users to send text
messages, audio, video, images and stickers to each other safely26.
The Telegram app is a cloud-based messenger app for smartphones and laptops that focuses on security and speed.
Smartphone operating systems that can use the Telegram application are iPhone/iPad, Android, Windows Phone, and
notebooks with PC/Mac/Linux, macOS and Web-browser operating systems.27.
Telegram is a cloud-based application, which makes it easy for users to access one Telegram account from different
devices at the same time28.
Advantages of the Telegram Application:
1. Telegram is a free app and will continue to be free (there will be no ads or fees forever).
2. Telegram sends messages faster because it is cloud-based.
3. Telegram is lighter when run, smaller application size.
4. Telegram can be accessed from various devices simultaneously, including smartphones, tablets, computers,
laptops and others.
5. Telegram allows us to share photos, videos, files with a maximum size of 1.5 GB per file.
With this application, digital citizens can interact well. Telegram bot technology has the advantage of one of the
secret features of the Telegram bot being able to download photos from Instagram29.
As we know, Instagram doesn't allow users to download photos directly from their app. Telegram bots can also be
used to practically create and develop online-based learning media30.
070001-3
Digital Learning
Online learning, in general, is learning that is done electronically using computer-based media and a network.
Online learning is also known as electronic learning, e-Learning, Digital learning, internet-enabled learning, virtual
learning, or web-based learning31. Online learning or e-Learning has started around 1970 is the result of learning
delivered online electronically using computers and computer-based media. The material is often accessed via a
network. Sources can come from websites and the internet.
In addition to providing instructions, e-learning can also monitor student performance and report student
progress32. E-learning access information and guides students to achieve specific learning outcomes. The term e-
learning has various meanings, whereas the current use of e-learning has two types, namely synchronous and
asynchronous. Synchronous means at the same time, where the learning process occurs at the same time between
educators and students. It allows direct interaction between educators and learners online. At the same time,
Asynchronous means not at the same time where students can take different learning times with educators providing
material. Asynchronous training is prevalent in e-learning because students can access learning materials anywhere
and anytime. Students can carry out learning and complete it at any time according to a predetermined schedule.
Education can take the form of reading, animation, simulation, educational games, tests, quizzes and task collection.33.
E-learning is an innovation that can utilize in the learning process, not only in the delivery of learning materials
but also in changes in the abilities of various student competencies34. Through e-learning, students listen to material
descriptions from educators and actively observe, perform, demonstrate, and so on. Teaching material materials can
be virtualized in various formats so that they are more exciting and more dynamic to motivate students to go further
in the learning process35. The application of e-learning for online learning today is straightforward by utilizing the
Learning Management System module, which is easy to install and manage like Moodle36.
Online learning was first recognized because of the influence of electronic-based learning through computer-based
learning systems37. Digital learning is a system that can facilitate students to learn more broadly, more, and varied 38.
Through the facilities provided by the system, students can learn anytime and anywhere without being limited by
distance, space and time. The learning materials studied are more varied, not only in verbal form but also in more
varied forms such as visual, audio, and motion. In general, online learning is very different from conventional learning.
Online learning emphasizes students' thoroughness and foresight in receiving and processing the information
presented online39. Digital learning also requires students and teachers to communicate interactively using information
and communication technology, such as computer media with the internet, telephone or fax. The use of this media
depends on the structure of the learning material and the types of communication required. Conversation transcripts,
examples of information, and written documents linking digital learning or learning via the Web showing full-text
examples are typical ways to document necessary learning materials online40.
METHODOLOGY
This type of research is qualitative research, a study that analyzes or describes a phenomenon or problem, either
in the form of events or social activities. The researcher uses a case study approach and literature review. The source
of the data in this study is the subject of valid and relevant data in the field. The researchers took samples using
purposive sampling and snowball sampling techniques, with a class population of three classes with a total of sixty
students with a class sample majoring in informatics. Mulia University, Indonesia.
Data collection techniques are observation, documentation and interviews, and supporting or secondary data by
looking for the latest sources, either in books, journals, news, articles, websites or desktops. The data analysis uses
data reduction, data presentation, and concluding. After that, check the validity of the data obtained by triangulation,
negative case analysis using reference materials.
Meanwhile, for data analysis, several stages can be carried out, including the following picture:
070001-4
Start
Problem analysis
System Procedure Analysis
Telegram
Program listing
End
FIGURE 1. Research stages
Problem analysis is carried out to find out what problems occur or may occur in the system; in this study, the
problem analysis is as follows:
The implementation of learning which is usually carried out conventionally using boards and paper, is considered
less effective because it requires a lot of time to make observations and corrections to assignments and student
learning; this implementation method also requires quite a lot of costs. The role of technology, e-learning based on
telegram media, is very important to provide solutions to the problems above.
FIGURE 2. Telegram Web
DISCUSSION
In this section, the research results on the stages of learning using the telegram application will be described. The
following are the results and discussions in this study. During the COVID-19 pandemic, various efforts were made to
reduce and prevent the spread of COVID-19, starting from social distancing, physical distancing to the implementation
of the lockdown. This application is also applied to educational activities, and learning activities must continue in
situations that do not allow face-to-face meetings. The majority of alternative solutions chosen by lecturers in learning
activities are to use teleconference with the hope that lecturers can teach as usual by conducting face-to-face
interactions but still utilising the distance principle. Lecturers can also share material in the form of power points
presented as regular. Learning activities using teleconference can replace learning activities carried out as face-to-face
activities, but not face to face. The following is an example of a virtual meeting using the Telegram application.
070001-5
FIGURE 3. Telegram learning online.
One of the advantages of the telegram application is the pollr feature that can be used to attend students, which is
not available in teleconference applications such as zoom, Webex and Skype, nor is it even available in telegram-like
applications such as WhatsApp. As lectures in general, lecturers will attend to students during lessons. In online
learning, absent activities must also be carried out, student attendance activities can be carried out with the poll feature,
the difference between poll and pollr is that polls can be timed by stopping them according to the time desired by the
lecturer and not available on telegram desktop. While the pollr feature is available on smartphones and telegram
desktops, in addition to the pollr quality, it looks more specific. There are column choices made by lecturers such as
entry, late, permission, and illness.
Based on the results of the analysis that has been carried out on the answers to the questionnaire, the score obtained
by the response is 100, while the maximum score is 100. Hence, the percentage obtained from the reliability test results
is 80.1%. So that the percentage value of eligibility of 80.1% is in the range of 61 to 80%, which shows that the results
of measuring the usability of e-learning telegram-based exams are "good".
TABLE 1. Usability level percentage
No Percentage Usability level

1 0%-20% Very bad
2 21%-40% Bad
3 41%-60% Enough
4 61%-80% Good
5 81%-100% Very good
CONCLUSION
After conducting several stages of testing and analysis of the problems, design, and implementation, some
conclusions can be drawn: the system was built with several steps, including problem identification, literature study,
design, implementation, and testing. In the manufacturing process, this system uses the telegram application and uses
the telegram bot. This system can conduct virtual meetings and process student exam data by displaying test data on
the telegram bot. Telegram bot fully supports the creation of an exam system or e-learning exam based on Telegram
bots quickly and efficiently.
ACKNOWLEDGMENTS
Mulia University supported this research; we also thank the research institute of Mulia University and the entire
academic community and all research colleagues and relatives for the assistance and cooperation that have been
established to make the research complete on time.
070001-6
REFERENCES
1. M. Ula, A. Pratama, Y. Asbar, W. Fuadi, R. Fajri, and R. Hardi, in J. Phys. Conf. Ser. (2021).
2. R. Hardi, Gunawan, and Sumardi, SNITT- Politek. Negeri Balikpapan 2018 (2018).
3. F. Fathiah, Nelli Raharti, and Muhammad Rizal Fachri, MULTICA Sci. Technol. 1, (2021).
4. Gunawan, Sumardi, R. Hardi, Suprijadi, and Y. Servanda, in J. Phys. Conf. Ser. (2021).
5. J.F. Rusdi, S. Salam, N.A. Abu, T.G. Baktina, R. Gumilar Hadiningrat, B. Sunaryo, A. Rusmartiana, W. Nashihuddin, P.
Fannya, F. Laurenty, N.M. Shanono, R. Hardi, S. Kuswayati, S.E. Damayanti, and S. Rahmawati, in J. Phys. Conf. Ser.
(2021).
7. Muhammad Yani, Fahru Rizal, and Muhammad Reza Padillah, MULTICA Sci. Technol. 1, (2021).
8. A. Junaidi, I. Kresna A, and R. Hardi, in J. Phys. Conf. Ser. (2021).
9. R. Hardi, JST (Jurnal Sains Ter. 3, (2017).
10. Y. Hendriana and R. Hardi, in 2016 Int. Conf. Inf. Technol. Syst. Innov. ICITSI 2016 - Proc. (2017).
11. V.C. Güngör, D. Sahin, T. Kocak, S. Ergüt, C. Buccella, C. Cecati, and G.P. Hancke, IEEE Trans. Ind. Informatics 7,
(2011).
12. S. Supriya, M. Magheshwari, S. Sree Udhyalakshmi, R. Subhashini, and Musthafa, Int. J. Appl. Eng. Res. 10, (2015).
13. R. Gouvea, D. Kapelianis, and S. Kassicieh, Technol. Forecast. Soc. Change 130, (2018).
14. C.G. Bletscher, J. Int. Migr. Integr. 21, (2020).
15. P. Faisal and Z. Kisman, Int. J. Eng. Bus. Manag. 12, (2020).
16. A. Muniasamy and A. Alasiry, Int. J. Emerg. Technol. Learn. 15, (2020).
17. M.Z. Iqbal, H.I. Alradhi, A.S.A. Alhumaidi, K.H. Alshaikh, A.M. AlObaid, M.T. Alhashim, and M.H. AlSheikh, Acta
Inform. Medica 28, (2020).
18. A. Alsharhan, S. Salloum, and K. Shaalan, Adv. Sci. Technol. Eng. Syst. 6, (2020).
19. M. Ivanova, Informatics Educ. 19, (2020).
20. A.M. Müller, C. Goh, L.Z. Lim, and X. Gao, Educ. Sci. 11, (2021).
21. S. Khadija Alhumaid, Multicult. Educ. 6, (2020).
22. M. Blankers, D. Van Der Gouwe, L. Stegemann, and L. Smit-Rigter, Eur. Addict. Res. 27, (2021).
23. L.H.X. Ng and J.Y. Loke, IEEE Internet Comput. 25, (2021).
24. L. Kusuma and T. Suwartono, Premise J. English Educ. Appl. Linguist. 10, (2021).
25. A. Martin-Gorgojo, J. Bernabeu-Wittel, M. Linares-Barrios, F. Russo-de la Torre, I. García-Doval, and E. del Río-de la
Torre, Actas Dermosifiliogr. 111, (2020).
26. O.G. Solomon, Acta Inform. Malaysia 5, (2021).
27. K. Suryati and I.G. Adnyana, JTAM (Jurnal Teor. dan Apl. Mat. 4, (2020).
28. Z. H. Iksan and S. Mohd Saufian, IJPTE Int. J. Pedagog. Teach. Educ. 1, (2017).
29. I.P.G.A. Sudiatmika, KERNEL J. Ris. Inov. Bid. Inform. dan Pendidik. Inform. 1, (2021).
30. S. Vahdat, Appl. Linguist. Res. J. (2020).
31. M.J. Sousa and Á. Rocha, Futur. Gener. Comput. Syst. 91, (2019).
32. N.B. Ødegaard, H.T. Myrhaug, T. Dahl-Michelsen, and Y. Røe, BMC Med. Educ. 21, (2021).
33. K. Sarnok, P. Wannapiroon, and P. Nilsook, Int. J. Inf. Educ. Technol. 9, (2019).
34. E.M. Mulenga and J.M. Marbán, Contemp. Educ. Technol. 12, (2020).
35. A. Lohr, M. Stadler, F. Schultz-Pernice, O. Chernikova, M. Sailer, F. Fischer, and M. Sailer, Comput. Human Behav. 119,
2021).
36. T.Y. Aikina and L.M. Bolsunovskaya, Int. J. Emerg. Technol. Learn. 15, (2020).
37. M.P.A. Murphy, Contemp. Secur. Policy 41, (2020).
38. A.D. Alharthi, M. Spichkova, and M. Hamilton, Requir. Eng. 24, (2019).
39. L. Ng, K.C. Seow, L. Macdonald, C. Correia, A. Reubenson, P. Gardner, A.L. Spence, S. Bunzli, and B.I.R. De Oliveira,
Phys. Ther. 101, (2021).
40. A.M. Syed, S. Ahmad, A. Alaraifi, and W. Rafi, Educ. Inf. Technol. 26, (2021).
070001-7
Improving Academic Creativity and Community Using
Google Apps for Education to Construct a Virtual Team
Sumardi1, a), Suhartati1, b), Nurfalah Setiawan1,c), Richki Hardi1,d), Hanafi 2,e)
Author Affiliations
1
Department of Informatics, Universitas Mulia, Balikpapan, Indonesia
2
Department of Informatics, Universitas Amikom Yogyakarta, Yogyakarta, Indonesia
Author Emails
a)
sumardi@universitasmulia.ac.id
b)
suhartati@universitasmulia.ac.id
c)
nurfalah@universitasmulia.ac.id
d)
e)
hanafi@amikom.ac.id
Abstract. The purpose of this research is to increase academic creativity in carrying out all forms of learning at
universities and form an educational community by establishing virtual team relationships to achieve the digital literacy
era and describe the use of Google Apps in the digital literacy era. The research method uses descriptive qualitative
research with data collection techniques using library research. Digital literacy is the ability to use information
technology from digital devices effectively and efficiently in various contexts of everyday life. One of the digital literacy
efforts is using information technology using Google Apps for education which the Google search engine company has
provided for all components of education in universities. By utilizing Google Apps in the learning process that opened
via smartphones and tablets and computers, it allows lecturers to prepare and present their learning materials online and
offline that are easily accessible by students.
INTRODUCTION
Entering the industrial era 4.0, marked by the rapid development of information technology, the impact of the
abundance of various information resources obtained digitally is unlimited1. As in the lives of today's millennial
generations, almost all aspects of their lives rely on the digital era, otherwise known as digital natives. Conditions
like this are also not surprising to change student behaviour in utilizing and managing information. The diversity of
forms and types of information should positively encourage students to be more selective and able to maximize the
use of information technology. Responding to the rapid advances in information technology, one of the international
organizations in the field of education, namely UNESCO, proclaimed the slogan literacy for all to improve the
quality of life in various lines of life, both individually, in families and society with the hope that it will provide a
multiplier effect in efforts to eradicate poverty, reduce mortality in birth control, to inhibit population growth,
achieving gender equality, sustainable development and peace.
Literacy is a human right and the basis for lifelong learning, covering various aspects of life. Literally, literacy is
defined as a set of reading, writing, and arithmetic2. Literacy is the ability to access and understand intelligently
reading, viewing, listening, writing or speaking activities. On the other hand, students can carry out critical analysis
in high-grade literacy competence, such as interviews, observations, and writing reports. This statement concluded
that the focus of literacy in question is the ability to read and write.
070002-1
Literacy is at the heart of students' ability to learn in school. Therefore, to face the challenges of the times and
information technology, in the 2013 curriculum, a school literacy movement was formed which places students as
subjects in learning, not objects3. The teacher's job is to provide a learning experience (facilitator). Teachers are no
longer positioned as the only source of learning but only as one of all learning resources that students can use.
Learning in the 2013 curriculum requires students to be active, starting from locating the required information
sources, sorting and selecting information according to their needs, and processing with reason and thoughts to
obtain the necessary conclusions and decisions. This study aims to describe the use of Google Apps in the era of
digital literacy for students.
The development of information and communication technology greatly affects all aspects of life in society,
including education4. Information and communication technology in education can play a dual role, namely as
media or learning aids and as learning materials or materials for students. Based on this, education is very closely
related and plays an important role in developing information and communication technology5. The direction of
information and communication technology development is currently more focused on computerized technology,
involving the internet and all-digital networks 6. For this reason, education must also transform from conventional to
digital information and communication technology 7. If education does not follow this current, it is feared that it will
be left behind, and of course, it will impact the quality of education. The amount of material presented to students is
directly proportional to the time required. This means that the more material students must master, the more time the
teacher needs to deliver it. Meanwhile, the time allocated for school hours is minimal. For this reason, the use of
digital technology-based media is needed to facilitate these limitations 8.
Teachers of today should not waste products that have been made free by the software company. On the other
hand, teachers should make maximum use of all the facilities provided in various applications for education 9. This is
done to improve the quality of better learning. The availability of applications or software as learning media are
available widely on the market, both free and paid. As a teacher in today's era, it is necessary to use media or
applications to improve the quality of learning10. One of the responsive software companies and provides various
applications to accommodate the educational needs of today's era is Google. Google offers Google Apps For
Education, a free application provided for educational institutions with information, communication, and
collaboration technology. Google Apps For Education facilities or products include Gmail, Classroom, Drive, Docs,
Slides, Forms and Sheets..
One of the new learning models in education is the bold learning model, which is carried out through unreal time
with Google Classroom. In implementing the use of technology and applications in learning, there are many
problems including, many teachers and students have not fully mastered technology, lack of time for educators to
explain the material to students, many offer but little explanation of the material, the use of classroom applications is
still rarely applied in schools, even though the operation is relatively easy, the teacher has difficulty in assessing
students. This study analyses bold learning by using the Google Classroom online learning application as a learning
medium.
LITERATUR REVIEW
Some literature reviews are obtained in related research, including digital literacy and Google Apps For
Education.
Literasi Digital
Media literacy is a series of media literacy movements, namely: the media literacy movement is designed to
increase individual control over the media they use to send and receive messages11. Media literacy is seen as a skill
that can be developed and is in a continuum where we are not media literate in all situations, at all times and to all
media12. Based on this statement, it can see that media literacy is an effort made by individuals so that they are
aware of the various forms of messages conveyed by the media, and are useful in the process of analyzing from
various truth points of view, understanding, evaluating and also using media 13. The term digital literacy was first
revealed that digital literacy is the ability to use information technology from digital devices effectively and
efficiently in various contexts of everyday life. Digital literacy is derived from computer and information device
literacy14. Computer literacy developed in the 1980s when microcomputers were increasingly used in business and
070002-2
society. Meanwhile, information literacy spread widely in the 1990s when information was more easily compiled,
accessed, and disseminated through networked information technology.
Digital literacy is an individual's awareness, attitude, and ability to use digital tools and facilities appropriately to
identify, access, manage, integrate, evaluate, analyze and synthesize digital resources, build new knowledge, create
media expressions, and communicate with others. In certain life situations, to enable constructive social action, and
reflect on the process15. Digital literacy is the ability to create and share in different situations and forms to
collaborate, communicate more effectively and understand how and when digital technology is used properly in
creating the process16. Thus, it concluded that the characteristics of digital literacy do not only refer to operating
skills and using various information technology and communication technology devices (hardware and software
platforms) but also to the process of "reading" and "understanding" the content of technological devices. And the
process of “creating” and “writing” into new knowledge. Digital literacy skills are emphasized on a person's attitude
and awareness in using ICT tools to communicate, express in social activities and achieve goals in various situations
in the life of the individual concerned. Digital literacy skills will be better if they can be developed in real-life
situations and solve problems17. The ability of digital literacy makes a person able to transform activities through the
use of digital technology devices. Everyone must have awareness as digital literate people in life, work, and study.
Solving problems in everyday life is important for the digital native generation through mastering digital literacy18.
Solving life problems including, completion of school assignments19.
Digital literacy competence is reviewed in critical thinking aspects; critical thinking aspects in digital literacy are
fundamental because of the variety of information on the internet and the ease with which internet users create
informative content20. Digital literacy skills as thinking development, meaning awareness of thinking about the tasks
assigned to someone21. Critical thinking should be an important part of developing the information literacy stage at
critically evaluating information 22.
Today the biggest obstacle in creating a literacy atmosphere in schools actually comes from internal schools,
including the inadequate ability of teachers and school infrastructure related to literacy, policies on literacy
programs that have not been formulated in writing, and no digital literacy program in the library. School, so that
students are not trained in finding, tracing, processing, and evaluating information effectively and efficiently. The
level of literacy among students also impacts the plagiarism culture that often occurs in the school environment.
Therefore, teachers and all school components must have good information literacy skills to teach information
literacy skills to students. On the other hand, digital media literacy skills are fundamental, and our abilities and
expertise are fundamental in mass communication. This skill is not always easy to master, it is more difficult than
just turning on the computer, showing television or turning the pages of a magazine, but it is essential to learn and
do it. This ability is media literacy, the ability to effectively and efficiently understand and use various forms of
media communication 23. In addition, media literacy skills have several objectives, including the following:
x Aims to help consumers have sufficient knowledge and understanding of media content so that they can
control the influence of media in their lives.
x To protect vulnerable and vulnerable consumers against the impact of media penetration of new media
culture.
x The purpose of media literacy is to produce citizens who are “well informed”. It can assess media
content based on their knowledge and understanding of the media in question.
Google Apps for Education
Google Apps is an application-based service created by Google. According to Google representative Sundar
Pichai, Google Apps allows users to upload and access various files such as videos, photos, spreadsheets and
PDFs24. GoogleApps can be accessed from anywhere. At home, at work, or even when you're away from your
devices. The application, it's used on Mac computers and other brands. In addition, it is used on tablet computers
and mobile phones with the Android operating system. Not only that, even visually impaired users can access and
use layer readers. Google Apps has a google drive application with a storage capacity of five Gigabytes and can be
used for free. Google with these services can facilitate users to collaborate, create, store and share documents with
other users25. The existence of online document management services by Google provides great benefits for
education.
Google Apps for Education consists of:
070002-3
x Gmail: An email account for each student, teacher, staff with many facilities provided and the same
functions as a Regular Gmail account with a storage capacity of up to 20 GigaBytes, thus helping
students store and find information quickly and easily.
x Google Calendar: Students can organize planned activities and events shared with others.
x Google Talk: Students can call and message each other at any time, from anywhere in the world.
x Google Docs: Students can share document files, work on a document together with groups in one
school, and publish the results to many people.
x Google Sites: Easily launch a website for a class, team, or project.
x Google Drive: A service that is directly connected to google docs, so you can share files
simultaneously.
x Google Classroom: A mixed learning platform dedicated to every educational setting intended to solve
the difficulties of creating, sharing and classifying paperless assignments.
The feature support that Google Apps has makes this service have enormous potential opportunities for the world
of education, including:
1. For the teaching and learning process, the teacher can carry out teaching and learning activities in the
classroom through Google Apps in the following ways:
x Students are given broadcast material or some video. Students are asked to study independently, and
students are asked to respond to questions or statements.
x The teacher gives the lesson's topic, and students are asked to respond to both questions and appropriate
statements so that interactive discussions occur. Fellow students can share their opinions. From the
responses of students who enter, the teacher can assess student activity and the quality of student
opinions.
2. Record student activities; Google Apps can record student activities in class, including recording student
attendance, grades, and activities. Teachers can use a spreadsheet on Google Drive to record attendance and
student grades. To track student activity, the teacher can give assignments so that from these assignments, it
can be seen how creativity, activeness/participation, and student opinions are.
3. Learning Evaluation, Google Apps can evaluate education, including learning evaluation. Teachers can
evaluate by making online quizzes and then giving an assessment automatically using Google Form or
Google Classroom.
4. Teamwork, Google Apps supports teamwork using Google spreadsheets. Where teachers can give
assignments done in teams, and students who are members of the team can access and edit the same
document simultaneously. Thus students are trained to work in teams.
5. Stimulate students' creativity through Google Apps trained teachers to give their opinions. This can
stimulate students' creativity, where students can freely express their ideas and ideas. In addition, students
also learn from other people's ideas and learn to understand other people's opinions.
6. With a word/document processing service, Google Apps facilitates its users to create various documents,
either in the form of a general script or a script intended for special needs, for example, for learning
mathematics. The word processor on Google Apps can also write formulas or symbols.
7. Presentation service allows users to create presentation files for education or classroom learning.
8. Spreadsheets owned by Google Apps are functionally almost the same as other office applications.
Spreadsheets on Google Apps used for learning, including learning evaluations, educational surveys,
learning models.
9. Image services provide Google Apps users with the opportunity to create objects for learning. For example,
making mathematical objects in flat fields, building spaces, etc.
10. The google form service allows someone to create forms for educational purposes
070002-4
FIGURE 1. Google suite facilities (source: www.sciping.com)
Utilizing Google Apps in the world of education, especially learning in the classroom, allows teachers to prepare
and present online and offline learning materials that are easily accessible to students 26. Teachers can upload their
learning materials in various forms and formats such as documents, audio, video, etc. These materials can be viewed
directly or downloaded via smartphones, tablets, and/or students' computers. In addition, the use of equipment such
as smartphones and tablets has many advantages, including being easy to carry anywhere, more affordable than
computers, providing learning opportunities without space restrictions, easy access to information via wireless,
encouraging the development of digital literacy, providing learning opportunities. Freely. Google Apps facilitates its
users to do creative, collaborative, exploratory learning and optimize the advancement of information technology27.
FIGURE 2. Most popular google play app (source: Appfigures;Statista)
METHODOLOGY
The research method uses descriptive qualitative research with data collection techniques using a literature study
of the listening technique, where one of the listening techniques is the note-taking technique. The note-taking
technique is a data collection technique by using books, literature or library materials, then taking notes or quoting
experts' opinions to strengthen the theoretical basis for writing scientific articles.
DISCUSSION
The online learning process is considered a new challenge in the era of industrial revolution 4.0, especially in the
midst of the Covid-19 pandemic. Like it or not, online learning is implemented. In the midst of a pandemic, digital-
based learning models have been massively maximized in almost all of Indonesia. Even though this model has not
070002-5
yet fully reached the lower social strata in society, this is because basically, the online learning model also has
conditions that must be met, namely access to digital information.
As technology advances, more and more people, especially young people, use gadgets, both notebooks and
smartphones. Especially before the industrial revolution 4.0, gadgets were increasingly for work and other activities.
To work usually on distance learning activities. Distance learning is a way of delivering education or training
programs remotely. This method does not require simultaneous interaction between the teacher/instructor and the
learner. Much distance learning has been implemented to date, especially since the Covid-19 pandemic has required
students to study at home. Several Google-based applications are used to support internet-connected learning.
Internet is not limited by distance and time. This is what makes learning be done anytime and anywhere. One of the
complete online learning methods is to use Google-based application learning media.
In online learning by teachers, at least the teacher must prepare hardware, such as laptops, earphones/headsets,
cellphones, and software, such as web browsers and compatible applications. In addition, it is necessary to prepare
an internet connection following video meeting standards and subject matter in the form of soft files or other forms
according to the planned meeting. The condition of the Covid-19 pandemic has made research activities also carried
out online, namely by training using video conferencing. Participants are also guided by the module and then guided
to practice continuously online, and then if there are questions, they submit them via email.
The training participants seemed enthusiastic because maybe this, like it or not, really needed them for learning
activities in the current pandemic conditions. The participants practice continuously based on the explanation results
during the video conference and based on the guidelines in the module. After that, if there are questions or
difficulties while practising, they can be asked via chat. Before all participants practised using the Google
application taught, the previous teacher participants were given a questionnaire to determine the problems before
being given training.
FIGURE 3. Communication through the classroom
The questionnaire results show that Google classroom is the second most used application after a google search,
which teachers most often use. The google classroom application, an online class, is an online learning media that is
very flexible both in place and time28. The advantage of using the Google Classroom application is that it is
considered more effective in delivering material and collecting assignments because it does not use stationery. Every
teacher can easily upload power points, videos and questions using Google Classroom. The results of other studies
conclude that the competence of teachers in carrying out learning improved by using the google classroom
application because it is more effective29. The effectiveness of google classroom can be seen from various aspects,
namely aspects of learning planning, aspects of designing and producing materials, aspects of delivery or learning
delivery methods; aspects of learning interaction; aspects of the evaluation of the implementation of learning, and
the criteria for the implementation of learning30. Because of the frequent use of these familiar Google applications,
they finally get used to it, and it becomes easy31. However, the results of this questionnaire indicate that teachers still
need to be given training related to the socialization of the use of all Google applications..
070002-6
Regarding anyone who uses the Google application, all of them answered that they use the Google application in
their family, although at least they do it themselves. They use the Google application while working, such as online
teaching, training, giving assignments, making teaching materials/materials, distributing questionnaires, processing
grades, opening/sending files, browsing/finding teaching materials, and assigning google forms. But some use the
Google application when travelling, looking for locations, communicating with friends/family, looking for
addresses, and looking for new things/fun and hobbies. With so many activities done through the Google
application, it is not surprising that teachers always use the Google application every day. When asked whether they
did not use the Google application in a day, more than 50% answered yes, but that was because when they were
busy with other work, it was like recapitulating grades. Some answered that they never used the Google application
in a day because they did not have an internet package/quota. Because of signal disturbances, it concluded that
teachers generally use the Google application if there is an opportunity either while at work or not at work to make it
easier to fulfil their needs. Everyday. The importance of the Google application for teachers is following their needs,
as many as 55% of the teachers use the Google application while working, both in the learning process and in
completing work assignments such as giving assignments to students, making materials, making questionnaires,
looking for teaching materials, sending files, and looking for references..
CONCLUSION
From the questionnaire results before the training, it concluded that all participants could understand the
applications on Google theoretically. Still, in practice, only the familiar ones are used, such as google mail, google
form and google drive. After introducing the functions and advantages of Google Classroom, making this
application from Google a choice for many teachers in schools as their main choice in the midst of the current
corona pandemic because the application on Google is quite helpful and represents learning like face-to-face at
school. It is proven to increase academic and community creativity using Google Apps for Education to build virtual
teams..
ACKNOWLEDGMENTS
The leadership of Mulia University supported this research; we also thank the research and community service
institutes of Mulia University and Amikom University Yogyakarta, and all research colleagues and relatives for their
assistance and cooperation in research activities.
REFERENCES
2. R. Hardi, Gunawan, and Sumardi, SNITT- Politek. Negeri Balikpapan 2018 (2018).
3. S. Sigalayan, M. Adriyanto, R. Hardi, J.T. Informatika, and K. Balikpapan, (2018).
6. J.F. Rusdi, S. Salam, N.A. Abu, B. Sunaryo, N. Agustina, H. Gusdevi, Hermansah, A. Hadhiwibowo, T.
Heryati, R. Hardi, R. Taufiq, M.T. Jufri, I. Sudarsono, E.W. Utomo, F.R. Kodong, and P. Fannya, in J. Phys.
Conf. Ser. (2021).
8. J.F. Rusdi, S. Salam, N.A. Abu, T.G. Baktina, R. Gumilar Hadiningrat, B. Sunaryo, A. Rusmartiana, W.
Nashihuddin, P. Fannya, F. Laurenty, N.M. Shanono, R. Hardi, S. Kuswayati, S.E. Damayanti, and S.
Rahmawati, in J. Phys. Conf. Ser. (2021).
9. R. Hardi, Preface 1807, (2021).
10. Z. Zaini, MULTICA Sci. Technol. 1, (2021).
12. J. McDougall, M. Readman, and P. Wilkinson, Learn. Media Technol. 43, (2018).
13. H. Park, H.S. Kim, and H.W. Park, J. Data Inf. Sci. 6, (2021).
14. P. Reddy, B. Sharma, and K. Chaudhary, Digital Literacy: A Review of Literature 11, (2020).
15. E.P. Yildiz, Cypriot J. Educ. Sci. 15, (2020).
16. M. Leaning, Media Commun. 7, (2019).
070002-7
17. A.D. Yazon, K. Ang-Manaig, C.A.C. Buama, and J.F.B. Tesoro, Univers. J. Educ. Res. 7, (2019).
18. A. List, Comput. Educ. 138, (2019).
19. D. Radovanović, C. Holst, S. Banerjee Belur, R. Srivastava, G. Vivien Houngbonon, E. Le Quentrec, J. Miliza,
A.S. Winkler, and J. Noll, Soc. Incl. 8, (2020).
20. S.S. Oh, K.A. Kim, M. Kim, J. Oh, S.H. Chu, and J.Y. Choi, Measurement of Digital Literacy among Older
Adults: Systematic Review 23, (2021).
21. T. Shepherd and M. Henderson, Can. J. Commun. 44, (2019).
22. K. Kumpulainen, H. Sairanen, and A. Nordström, J. Early Child. Lit. 20, (2020).
23. E.F. Rusydiyah, E. Purwati, and A. Prabowo, Cakrawala Pendidik. 39, (2020).
24. A. Daud, Indones. J. Econ. Soc. Humanit. 1, (2019).
25. M.E. Brown and D.L. Hocutt, J. Usability Stud. 10, (2015).
26. J.M. Locketz, ProQuest Diss. Theses (2019).
27. E. Railean, Int. J. Comput. Sci. Res. Appl. Int. J. Comput. Sci. Res. Appl. ISSN 02, (2012).
28. K. Ansong-Gyimah, Int. J. Emerg. Technol. Learn. 15, (2020).
29. M.L. Hidayat, W.H. Prasetiyo, and J. Wantoro, Humanit. Soc. Sci. Rev. 7, (2019).
30. J.A. Kumar and B. Bervell, Educ. Inf. Technol. 24, (2019).
31. R. Hardi, 12, (2015).
070002-8
Increasing the Importance of Digital Technology as a
Technopreneurship Media in Higher Education
Richki Hardi1,a), Agung Sakti Pribadi2,b), Mundzir1,c), Agustinus
Noertjahyana3,d), Jack Febrian Rusdi4, e)
Author Affiliations
1
Departement of Informatics, Universitas Mulia, Balikpapan, Indonesia
2
Departement of Law, Universitas Mulia, Balikpapan, Indonesia
3
Petra Christian University, Surabaya, Indonesia
4
Departement of Informatics, Sekolah Tinggi Teknologi Bandung, Bandung, Indonesia.
Author Emails
a)
Corresponding author:richki@universitasmulia.ac.id
b)
agungsakti@universitasmulia.ac.id
c)
mundzir@universitasmulia.ac.id
d)
agust@petra.ac.id
e)
Abstract. In overcoming employment problems, continuous efforts are needed to reduce the disturbances felt from year to
year, which is increasing, especially from college graduates. Technopreneurship course material is expected to help and
provide alternative solutions, especially in seeing the changing mindset. Job seekers become job creators.
Technopreneuship contains the theory and practice of entrepreneurship by delivering innovation through the application of
technology owned by universities in particular and other technological resources to create innovative products to give birth
to new technology-based entrepreneurs. To make it happen, it is necessary to know the conditions and obstacles of higher
education in mastering the science of technopreneurship. Because currently, new universities provide an introduction to
entrepreneurship, not yet in the form of technology entrepreneurship. For this reason, researchers have conducted studies
to provide ideas to the parties involved in the development of Technopreneurship related to conditions, model development,
and policies needed to encourage the development of Technopreneurship in universities, namely by increasing the
importance of internet technology as a medium for technopreneurship in universities.
INTRODUCTION
In the development of the digital era, many young technopreneurs are innovative and able to move the wheels of
the Indonesian economy for the better. Technopreneurs are entrepreneurs who utilize technology to produce promising
innovations for consumers. Technopreneurs run businesses differently from other entrepreneurs. A Technopreneur's
company has high growth potential and requires intellectual knowledge. So there is a strong relationship between
technology development, innovation and entrepreneurship. Currently, the problem of employment in Indonesia is
faced with an imbalance between job opportunities and labour supply. This gap has led to intense competition for jobs.
Even college graduates, it is not easy to compete in getting a job. Efforts to accelerate the growth of new businesses
are significant, especially in response to the increasing supply of labour from year to year. One of the efforts to
accelerate the development of new companies is to make breakthroughs by changing the perspective of college
graduates from job seekers to job creators.1. To realize this effort, universities have provided entrepreneurial training.
However, it is still more extensive in terms of theory in practice. Even so, this can already be a magnificent essential
capital. Furthermore, the entrepreneurship material needs to be added with technology skill content, namely the ability
070003-1
to innovate through the application of technology. Technology-based entrepreneurship education, known as
technopreneurship, is an effort to synergize theory and practice from various competencies in science related to
technology and industry2. Therefore, technopreneurship education can be used as a business atmosphere learning
process. Technopreneurship education in the future can be developed at universities in Indonesia. However, in its
development, universities are still facing various problems, both the absence of policies and the readiness of
universities for teaching staff and educational infrastructure.
The term "Technopreneurship" is currently being discussed a lot in various media, both social media, mass media,
and electronic media3. Technopreneurship is considered as a concept that is derived from "Entrepreneurship", which
both share the principle of seeking as much profit as possible but focuses more on a business that applies a certain
technology, not just the replication of other businesses4.
Technology plays an essential role in the development of the modern world today; the continuous emergence of
new technologies and the application of more and more technology requires continuous innovation so that the use of
technology can be effective and achieve its goals. Learning about technology requires support from human resources;
in this case, it can be studied at universities and requires practical work that is carried out regularly. Technology is a
way to process something. Cost and time efficiency occurs to produce quality products by considering market needs,
solutions to problems, application development, improvement of production effectiveness and efficiency and
modernization. A technopreneur is never enough to learn only one or two technologies but must be sensitive to
technological innovation, and creative ideas are needed to support it.5.
The various advances that have been achieved began with research and new findings in the field of technology or
inventions which were then developed in such a way as to provide benefits for the creators and the community of
users. The phenomenon of business development in technology begins with creative ideas in several research centres
that can be developed so that they have a selling value in the market. The initiators of ideas and product creators in
the technology field are often referred to as technopreneurs because they can combine their knowledge through product
creations/ideas with entrepreneurial skills through selling products produced in the market. Thus, technopreneurship
combines technology, the ability of science and technology with entrepreneurship working alone to bring profit
through business processes6.
Currently, business developments in technology are essentially the result of synergies between owners of creative
technopreneur ideas, who are generally affiliated with various university research centres, and capital providers to be
used in doing business. The relationship between these three elements has encouraged the development of technology
businesses in several countries, such as Silicon Valley in the United States, Bangalore in India, and several other
countries. In Indonesia, the synergy between the three parties has not been well developed. The development of various
innovation centres and business incubators in technology in several universities and research institutions is a positive
effort to build technopreneurs in Indonesia.
Departing from Community Needs Community needs are business opportunities. Especially if there are
community needs that any party in this world has not met, almost all technology-based products that are very well
known and widely purchased today are those that depart from people's needs.7. Cars, motorcycles, cell phones,
television, internet, cellular providers, social media, various electronic products, and gadgets originated from people's
needs. If you want to become a technopreneur, start from the needs and problems of the community so that we can
have specific ideas or ideas to provide solutions through technology that we can develop into a business core. This
also makes our products attractive to the public so that we can continue to develop them for the better8.
Enrich yourself with ideas and inspiration ideas and inspiration is the beginning of the emergence of a business
idea. It takes a brilliant idea to start a business and maintain it in this highly competitive era. The products we produce
do not need to be new but must be innovative by modifying something existing and making its function much better
or diverse. Ideas and inspiration can sometimes come by themselves, but the best way is to develop the concept and
inspiration itself. The trick is to enrich your knowledge by reading, attending seminars or workshops on
technopreneurship, or talking to technopreneurs directly9. Whether we realize it or not, these things will give rise to
an original idea that we can develop as a business, Plan carefully and do it quickly. A technopreneur must analyze the
market, design a product, make a marketing strategy, determine prices and target markets, develop an organizational
structure, and be responsible for all business processes. It is the ability that a technopreneur must have in general in
making a business plan. But of course, the program will not come true if it is not realized. So, start as soon as possible
or even now. Start with easy and simple things like finding inspiration, designing products or creating promotional
strategies.
Add Value to Products The products we produce can be precisely the same as other entrepreneurs. But one thing
that makes a specific product more preferable and more in demand than other similar products is value10. The value
that we can add to our products is, of course, diverse and following the innovation and creativity of each technopreneur.
070003-2
The matter described here is not about price but an added value. This will undoubtedly increase the selling value of
the product, especially for people who want games that are not only entertaining but also educational.
Entrepreneurship Technology concepts

concepts and theories and theories
Technopreneurship
Internet technology
development in
technopreneurship
FIGURE 1. Technopreneurship direction
LITERATUR REVIEW
The literature review section will be divided into three parts: the concept of entrepreneur, technopreneur, the
importance of digital media for technopreneur, and previous research related to this research.
Enterpreneur
Entrepreneurship is a value creation process by using specific resources to exploit opportunities11. The concept of
entrepreneurship has received extensive and intensive attention from various groups, both from experts and
practitioners such as economics, business management and bureaucrats working in the public sector. Several aspects
need to be considered in the process of implementing the concept of entrepreneurship in a government or government,
namely:
Customer choice, or customer choice, makes it easy for service users in choosing or determine service
providers according to the characteristics of their aspirations.
Budget efficiency, rampant corruption, collusion, and nepotism that occur in the field make the
government have to be more assertive in providing supervision regarding matters relating to these budget
allocations.
Innovation and creativity are an essential part of entrepreneurial-style governance. This innovation and
creativity is a tool that can be a solution to overcome the dynamic changes that occur in society.
Entrepreneurship Concept Development

The concept of the entrepreneur was introduced in the 18th century in France when an economist named Richard
Cantillon linked the burden of risk that the government and the caregivers must bear in running the economy.
In the current century, with technological advances and various changes, the world feels as if it has become narrow
and has lost its boundaries. Along with this fact, it must be acknowledged that the progress and changes that have
occurred are evidence of several multinational entrepreneurs from various parts of the world. They are present as
agents of change, and they are born with several innovative ideas for the development of the business world and
economic development in general.
Technopreneurship
An entrepreneur establishes a business by identifying existing opportunities and combining the necessary resources
despite having to take risks and uncertainties to gain profit and growth. 12. Entrepreneurship has become a benchmark
for the economy, especially creating jobs and prosperity for people in developed and developing countries. Meanwhile,
according to Presidential Instruction No. 5 of 1995 explains that entrepreneurship is the spirit, attitude, behaviour, and
ability of a person in handling a business or activity that leads to seeking, creating, implementing new ways of
070003-3
working, technology, and products by increasing efficiency to provide services. Better and make more profit13.
Entrepreneurship is an effort to create added value by combining resources through new and different ways to increase
competition.
Technopreneurship is a collaboration between technology and an independent business spirit with the spirit of
building a business to generate jobs and build the Indonesian economy and technology. Technopreneurship is one of
the processes of creating a new company using technology as its basis to become an incubator to achieve success14.
Technopreneurship development requires the concept of entrepreneurship, business skills, marketing, business
plans, management or business, technology skills invention, innovation, supply and demand for technology,
intellectual property management, and product or packaging design.
Digital Media for Technopreneur
The use of the Internet in business has grown, from the electronic exchange of information to applying business
strategies, such as marketing, sales, and customer service. Due to the Internet, companies' marketing, products and
services have become an interactive process today15.
Facing the digital era, people have started to be creative in creating new technology-based business models. This
model is called technopreneur, a form of entrepreneur, who both have an essential role for economic development in
the future16.
The Internet of things (IoT) describes the network of physical objects—"things"—that are embedded with sensors,
software, and other technologies to connect and exchange data with other devices and systems over the Internet. In
the future, hundreds of billions of intelligent sensors and devices will interact with one another without human
intervention on a Machine-to-Machine (M2M) basis17.
FIGURE 2. M2M application examples, different rows reflect different human wishes, different columns reflect who pay for the
service/applications
METHODOLOGY
The method used is by conducting participatory observations, in-depth interviews, and documentation of data
sources. The data sources of this research are divided into two, namely primary and secondary data sources. Primary
data sources were obtained from the results of research respondents, including university leaders, faculties,
departments/ study programs, entrepreneurial activity support units, and entrepreneurship lecturers. Data was collected
through interviews, observation, and documentation. Interviews were used to collect data on policies related to
entrepreneurship education and their implementation; words to collect data on the performance of lecturers and
entrepreneurship support activities, and documentation is used to collect data about the results obtained from
entrepreneurship education activities. The research data were analyzed using descriptive analysis techniques, both
quantitative and qualitative and also ratio analysis techniques.
DISCUSSION
The importance of technopreneurship today is related to its attachment to science and technology. When a country
uses an approach to increasing technological capabilities as a driver of increasing national production and in many
countries as a competitive advantage strategy, then technopreneurship is a program that is included as an integral part
of increasing the entrepreneurial culture. Technopreneurship needs to collaborate on culture and conceptions, namely
070003-4
the culture of innovation, entrepreneurship, and creativity, as well as the concept of business incubators, research,
development, knowledge management and learning organizations, which are supported by their own entrepreneurial
capabilities, connectivity and collaboration.
Universities have an important role in giving birth to a young generation with high character and competitiveness
that will come in the midst of the challenges and changes facing the world community today. high adaptability,
competitiveness and entrepreneurial spirit are the main prerequisites that need to be possessed by today's young
generation in order to become job creators or technopreneurs.
Talents from competitive universities will greatly support the development of the business world. These talents
will also play an important role in the digital economy transformation process.
most of the discourse in our country directs Technopreneurship as in the second definition. Especially in this day
and age, business via the Internet is growing. There is a belief that Technopreneurship is a business solution in times
like this.
Education has an important role in the development of the whole person and the development of Indonesian society
as a whole. Human development must be carried out as a whole, which includes the development of thinking power,
heart power, physical strength, and mastery of science, technology, art and sports18. In addition, human development
is also expected to produce people who are capable and able to play an active role in building Indonesian society as a
whole.
The task of the education sector, both formal and informal, is not only to produce educated humans, but more
broadly, the education sector must be able to create independent human beings. With the fact that not all Indonesians
of productive age and classified as labour force can be absorbed in the world of work, then the education sector is
responsible for finding solutions, how so that the output produced is not only oriented to become workers, in addition
to the role of the education sector to introduce and motivate their students to understand that apart from being a worker,
entrepreneurship is also a promising field to explore.19.
To produce successful young entrepreneurs requires sincerity and seriousness from universities in carrying out the
entrepreneurial campus mission. Entrepreneurship programs initiated and run by various universities, especially in
Indonesia, should be used as role models in starting to focus universities on producing successful young entrepreneurs
who can integrate entrepreneurial concepts and technology concepts to increase their business development potential.
The development of national entrepreneurship is a significant and noble task that requires the togetherness of all
components of the nation. The growth of new entrepreneurs cannot be done partially or by one agency because each
agency has limitations according to their respective main tasks and functions. The program for the growth of new
entrepreneurs must be carried out comprehensively by involving all relevant agencies, both central and regional
governments, educational institutions, business entities and non-governmental organizations. The spirit of
togetherness and synergy of elements of the government, academia, the business world, new entrepreneurs and all
other components of society need to be continuously encouraged so that more Indonesian children choose their
profession to become entrepreneurs.
In the context of developing a more effective national entrepreneurship, it is necessary to consider establishing a
national entrepreneurship development coordinating institution that maintains the synergy and togetherness aspect of
all components of the nation by providing access to more structured coordination in the world of education, technology
and creative bodies to accommodate the results of technopreneurship20.
CONCLUSION
Entrepreneurship education in higher education is related to building entrepreneurial character, entrepreneurial
mindset, and entrepreneurial behaviour that is always creative and innovative, creates added value or good values,
takes advantage of opportunities and dares to take risks. Facing the challenges of a highly competitive future,
entrepreneurial behaviour is needed for all fields of work or profession. Therefore, entrepreneurship education can be
implemented in universities and applied to all students regardless of the knowledge studied because entrepreneurship
education is not business education.
ACKNOWLEDGMENTS
We would like to thank the leadership of the University of Mulia, LP3M University of Mulia, colleagues and
researchers from the University of Mulia and external researchers who have supported this research. Hopefully, this
research can be helpful for universities local areas and become a reference for other researchers.
070003-5
REFERENCES
1. W.L. Koe, M.H. Mahphoth, N.E. Alias, R. Krishnan, and A.F. Arham, J. Educ. Soc. Res. 11, (2021).
2. M. Bomani, G. Gamariel, and J. Juana, J. Gov. Regul. 10, (2021).
3. H. Hidayat, S. Herawati, E. Syahmaidi, A. Hidayati, and Z. Ardi, Int. J. Eng. Technol. 7, (2018).
4. M. Selladurai, SELP J. Soc. Sci. VII, (2016).
5. A.A. Abbas, Int. J. Sci. Eng. Res. 9, (2018).
6. R.A. Fowosire, O. Elijah, and R. Fowosire, Type Double Blind Peer Rev. Int. Res. J. Publ. Glob. Journals Inc 17, (2017).
7. H. Nurdiyanto, Glob. J. Eng. Educ. 20, (2018).
8. H. Hidayat, B.Y. Tamin, S. Herawati, A. Hidayati, and A.P. Muji, Int. J. Innov. Technol. Explor. Eng. 8, (2019).
9. A.D. Amante and T.A. Ronquillo, Australas. J. Eng. Educ. 22, (2017).
10. B.A. Soomro and N. Shah, World J. Entrep. Manag. Sustain. Dev. 17, (2020).
11. E. Ismail, S. Samsudi, and D. Widjanarko, J. Vocat. Career Educ. 2, (2017).
12. N. Haryanti, M. Nor, S. Maziah, A. Rahman, Y. Hayati, N.M. Naziman, S. Norbaya, M. Rashid, N. Farleena, and M.
Aznan, Adv. Sci. Lett. 23, (2017).
13. S. Saad, A.S.M.M. Hoque, and Z. Awang, Proceeding Int. Semin. Entrep. Bus. 2019 (2019).
14. H. Hidayat, Z. Ardi, Yuliana, and S. Herawati, Int. J. Econ. Bus. Res. 18, (2019).
15. D. Games, R. Kartika, D.K. Sari, and A. Assariy, J. Sci. Technol. Policy Manag. 12, (2020).
17. J. Liu, M. Chen, and H. Liu, J. Data, Inf. Manag. 2, (2020).
(2021).
070003-6
The Use of E-Learning to Increase Student Innovation in
Technopreneurship
Vidy1, a), Richki Hardi1, b), Yamani1, c) and Wahyu Nur Alimyaningtias1, d)
Author Affiliations
1
Departement of Informatics, Universitas Mulia, Indonesia
Author Emails
a)
Corresponding author: vidy@universitasmulia.ac.id
b)
richki@universitasmulia.ac.id
c)
yamani@universitasmulia.ac.id
d)
wahyuutias@gmail.com
Abstract. Higher Education is one of the institutions that have an essential role in creating competitive and innovative
resources to produce various works or products that can encourage economic growth. The current pandemic situation has
no reason to be unproductive while at home, but technology can help daily activities. One form of technology utilization is
carried out through entrepreneurship-based student innovation activities carried out by the Higher Education Study Program
in the form of training and assistance to others to foster technopreneur interest through technology. The approach taken in
this program uses qualitative. A coaching clinic does mentor on the proposed business idea or product. The results of the
research are e-learning technology to increase student innovation in technopreneurship.
INTRODUCTION
Globalization has brought many changes, especially in technology and information1. Today's technology and
information have become an inseparable part of social life; even the development of technology and information has
dramatically influenced today's education. In the world of education, the use of technology and information is an
essential part of the learning process in the classroom, not only regarding the curriculum but also regarding learning
media.
Human resources play the most crucial role in the progress of a nation, so activities to increase the potential of
human resources are an essential priority that must be considered carefully. However, the large and dispersed number
of human resources has become a challenge for equal distribution of education in Indonesia to increase the potential
of human resources2.
Information technology in the world of education promises many things, especially in the academic world, through
the development and ease of communication. One example of technological progress in education is the concept of e-
learning or distance learning. Some experts even predict that permanent learning models in the future may not be
found anymore. The very basic idea of e-Learning is that the teacher and the participants are separated in terms of
place and time in the learning process. Because the teacher and the participants are in different areas and or at other
times, it is necessary to have an e-media that conveys information that guarantees interaction between them 3.
Conventional learning systems are considered no longer in accordance with the times, apart from due to a shift in
education that was previously focused on lecturers to focus on students and the surrounding environment where the
development of technology and information greatly affects. Many students feel that conventional learning systems are
070004-1
not able to attract and foster student interest and achievement 4. So that the presence of technology and information
becomes an important solution to foster student interest in learning. In connection with this phenomenon, the
sophistication of technology and information has presented e-learning-based learning media, where e-learning-based
learning media are expected to be able to increase student interest and learning achievement. E-learning-based learning
media is a tool to deliver learning materials for students, so as to stimulate interest and student learning achievement.
By understanding e-learning, it is hoped that students will become aware of technopreneurs and creativepreneurs.
Currently, educational institutions have focused on developing applications of technology and information in the
business sector and providing skilled human resources in managing business by utilizing technology and information
tools. In connection with this description, this study will discuss the Utilization of E-Learning to Increase Student
Innovation in Technopreneurship5.
LITERATUR REVIEW
There are four sections of discussion in the literature review including technopreneur, entrepreneur and eLearning.
Technopreneur
The business concept of technology and information is currently a trend in the business world. 6. The power of
product innovation and its business model can grow the business quickly and sustainably. The tendency to prioritize
business based on technology and information innovation occurs in various parts of the world, including Indonesia.
Many college graduates give birth to entrepreneurs who have the ability and understanding of technology and
information. This effort is getting stronger after the government declared this period as the year of Creative Indonesia.
An entrepreneur can add value to a product he sells by making the most of the resources he has. Technopreneurs can
provide more value to the products sold based on their knowledge and technology7.
Creativepreneur
Creativepreneurship is a business activity or business that utilizes creative ideas and involves the added value of
art and design as the backbone of its primary activities8. A creativepreneurship will always create innovations to
change something existing or create something new that is more valuable. Creativepreneurship can also be understood
as an attitude and way of thinking about creating creative work and selling well in the market. The creative economy
is a business activity of producing goods and services based on intellectual property by relying on expertise, talent,
and creativity. The creative industry is proven to contribute to the national economy to date. In addition to contributing
to the amount of Gross Domestic Product, creative industries are also able to absorb a lot of workers and affect
Indonesia's export performance. The current form of innovative business that is increasingly fluttering and taking part
in the industry are businesses related to the use of information technology or generally also referred to as a digital
business, which is a form of creative endeavour that is increasing because it is related to public entertainment in the
form of the entertainment business9.
E-Learning
e-Learning was introduced to the world of education starting in 1996. The term e-Learning can be defined as a
form of information technology applied in education in the form of virtual schools. The term e-Learning is more
accurately intended as an effort to make a transformation of the teaching and learning process in schools into digital
form which is bridged by internet technology. The focus of e-learning is more on the efficiency of the teaching-
learning process while the teaching methods and teaching materials can still refer to the national curriculum. Students
are more passive and positioned as consumers of knowledge. With e-Learning technology, all teaching and learning
processes that are usually found in a classroom are carried out live but virtual, meaning that at the same time a teacher
teaches in front of a computer in one place, while students follow the lesson from another computer. in a different
place. In this case, directly the teacher and students do not communicate with each other but indirectly they interact
with each other at the same time.
Understanding of electronic learning or e-learning often changes along with the development of technology and
information 10. In general, e-learning is teaching and learning that uses electronic circuits, networks and the internet
070004-2
to deliver learning materials. Computers, internet are some of the electronic media that fall into this category. E-
learning is a relatively new information technology to support the success of the teaching and learning process. The
teaching and learning process through e-learning is carried out using various information technology facilities, such
as computers both hardware and software, network technologies such as local area networks and wide area networks,
as well as telecommunications technology such as radio, telephone and satellite. students, improve students'
independent learning abilities and improve the quality of learning materials.
Students' interest in learning needs to get adequate attention. It stated that interest means a persistent tendency.
Someone interested in inactivity will pay attention to the movement consistently with pleasure. Interest is an
unintentional concentration of awareness born full of will and depends on talent and environment. The statement
explains that interest is a focus of attention. Interest is a sense of preference and a purpose of attachment to something
or activity without anyone telling. The statement identifies that interested people will be attracted. Being interested in
this is a form of feeling happy about something. Based on some of these opinions, it can be concluded that student
interest in learning is the tendency of students to be involved consistently and continuously in education. This interest
is built on attention and feelings of pleasure. Interest in learning in this study will be measured using the following
indicators:
Feeling happy.
Feelings of pleasure are subjective psychological symptoms associated with symptoms of knowing and being
experienced in the quality of pleasure or displeasure at various levels during the learning process.
Attention.
Attention is the concentration or concentration of all student activities aimed at learning materials and the learning
process.
Activity.
Activities related to the number of students expressing themselves, manifesting their feelings and thoughts in
spontaneous actions in the learning process.
Learning achievement
Learning achievement is the result obtained by students after following a specific learning process. Learning
achievement is the mastery of knowledge or skills developed by the subject matter, usually indicated by the lecturer's
test scores or scores. Based on this understanding, it can be understood that learning achievement is the result of
students' abilities in certain learning materials12. Assessment can be in the form of numbers or letters. Students' success
in achieving learning achievement is influenced by several factors: a good level of intelligence, lessons according to
their talents, high interest and Attention in learning, good motivation in education, suitable learning methods, and
learning strategies developed by lecturers. A family atmosphere that encourages children to progress, besides an
orderly, orderly, and disciplined school environment, is a driving force in achieving learning achievement.
METHODOLOGY
This method is generally defined as a process, method, or procedure used to solve a problem. The methods used
to collect data in this study are as follows:
Data Collection Method, Literature study is a data collection method directed to find data and information through
documents, both written documents, photos, pictures and electronic documents that can support the writing process.
This method refers to books, scientific journals and the internet. So it can be said that the literature study can affect
the credibility of the research results. Literature study is a data collection method that is directed at finding data and
information through documents, both written documents, photos, pictures and electronic documents that can support
the writing process. This method refers to books, scientific journals and the internet. So it can be said that the literature
study can affect the credibility of the research results. The following is a literature study related to e-learning to add
creativepreneurship innovation to students..
Research that describes an adaptive e-learning system contains learning styles that refer to the way students prefer
to learn and play an important role13. In its application, the system can offer valuable advice and instructions to students
and teachers to optimize the student learning process, increase student innovation in technopreneurship, increase
student innovation in technopreneurship. The integrated learning e-learning system called the blended learning system
that is applied is one form to improve the quality of education towards a modern and quality one.
This study discusses the benefits of technological developments used in e-learning methods that can support the
teaching and learning process and increase learning efficiency. It illustrates that e-learning contains instructions sent
on digital devices14. In learning, it can be used on computers, laptops and smartphones to support e-learning digital
070004-3
learning activities. E-learning is designed for asynchronous learning or self-study. E-Learning also overcomes the
problem of conventional education that is computerized 15.
DISCUSSION
The first stage of learning begins with providing business motivation. The challenge in management is that today's
universities must be open about their education, lecturers, and teaching. This causes all universities to compete, both
nationally and internationally. Each university must survive by upgrading all aspects of it. Especially learning and
teaching content adapted to industry developments. Because college graduates will go directly into society. The
Indonesian government encourages the improvement of the quality of graduates through many programs, one example
of which is the latest program by the Minister of Education and Culture. Universities are required to direct the quality
of their graduates to suit the needs of today's industry in Indonesia. By producing graduates to become
creativepreneurs, students can play a fundamental role in generating value in product innovation. This policy aims to
increase link and match with the business and industrial world and prepare students for the world of work from the
start. In its application, some things can be done, including internships, teaching in schools, humanitarian projects,
independent studies, entrepreneurship or entrepreneurs, research/research, student exchanges and projects in villages.
In entrepreneurial activities, students can develop themselves directly in society.16. Challenges of the learning process
Curriculum at tertiary institutions are structured and designed so that it becomes the basis for teaching and learning
activities that end with the success of students mastering knowledge, abilities, skills, and achieving predetermined
qualifications. Currently, universities have entrepreneurship courses as the initial stage of understanding
entrepreneurship17. Observations regarding the implementation of e-learning resulted in several cases that can be used
as examples of learning. If e-learning is applied purely, it means that it will produce positive and negative excesses
compared to conventional teaching. On the positive side, students will look for complementary knowledge by looking
for various other references and being more critical of what they face.
e-learning program
Entrepreneurship Innovative
Concept
Teaching and Quality Competence

learning activities
Technopreneurship
concept Competitiveness
E-learning
Program
FIGURE 1. E-Learning and Graduate Quality
CONCLUSION
An important finding in this research is technology in entrepreneurship learning. The technology used in e-
learning.18. Indicators of learning success are seen based on eight characteristics that affect the ability of
technopreneurs, which include optimization of technology and information, scientific implementation, self-
confidence, results-oriented, risk-taking, leadership, product originality, and future-oriented. Technical skills must be
improved to keep up with technological developments. Having entrepreneurial experience can increase students' self-
confidence in facing the world of work. Based on the analysis conducted, the authors conclude that the research
concept of e-learning supports student innovation in bident technopreneur. In its application, e-learning should use the
basis of entrepreneurship. In this way, the resulting creation can be applied directly by the technopreneurship
070004-4
community based on e-learning19. The creativepreneurship foundation will encourage students, lecturers, and
academics to be creative and innovative in improving the quality of services, education, teaching, and learning, further
enhancing graduates' quality, competence, and soft skills so that their competitiveness will automatically increase. E-
learning should be adjusted and balanced with conventional teaching to guard against the adverse effects of distance
learning. Universities are still a medium of learning and education for students 20.
ACKNOWLEDGMENTS
Thank you to the University of Mulia academics who have supported this research and colleagues who have
encouraged this research to be completed. LP3 M Mulia University fully supports this research.
REFERENCES
(2021).
3. Y. Hendriana and R. Hardi, in 2016 Int. Conf. Inf. Technol. Syst. Innov. ICITSI 2016 - Proc. (2017).
4. M. Ula, Mursyidah, Y. Hendriana, and R. Hardi, in 2016 Int. Conf. Inf. Technol. Syst. Innov. ICITSI 2016 - Proc. (2017).
6. T. Wibawa, H. Widjanarko, H.S. Utomo, and ..., … LPPM UPN “ … 1, (2020).
7. Joang Ipmawati and Y. Hendriana, MULTICA Sci. Technol. 1, (2021).
8. M. Mahamood, Int. Conf. One Asia Community 1, (2019).
9. M. Pujiyanto, R. Hidayat, and M. Sumarwahyudi, in (2018).
10. L. Mardiana, F.E. Nilawati, R. Zulfiningrum, and Z. Umami, in (2020).
11. A. Selvarani and K. Venusamy, Int. J. Manag. 6, (2015).
12. N.N. Okorie, D.Y. Kwa, S.O.O. Olusunle, A.O. Akinyanmi, and I.M. Momoh, Eur. Sci. Journal, ESJ 10, (2014).
13. R.A. Fowosire, O. Elijah, and R. Fowosire, Type Double Blind Peer Rev. Int. Res. J. Publ. Glob. Journals Inc 17, (2017).
14. H. Hidayat, B.Y. Tamin, S. Herawati, A. Hidayati, and A.P. Muji, Int. J. Innov. Technol. Explor. Eng. 8, (2019).
15. S. Suradi, R. M. Yasin, and M.S. Rasul, J. Tech. Educ. Train. 9, (2017).
16. S. Mashingaidze, Environ. Econ. 7, (2016).
17. B.A. Soomro and N. Shah, World J. Entrep. Manag. Sustain. Dev. 17, (2020).
18. A.D. Alharthi, M. Spichkova, and M. Hamilton, Requir. Eng. 24, (2019).
19. M. Selladurai, SELP J. Soc. Sci. VII, (2016).
20. M. Ivanova, Informatics Educ. 19, (2020).
070004-5
Design of Google Apps-Based Helpdesk Ticketing System at
University Using Extreme Programming Approach
Yusuf Wibisono1,a), Mada Aditia Wardhana1,b) and Wisnu Hera Pamungkas1,c)
Author Affiliations
1
Mulia University
Jl. ZA Maulani No. 9 Balikpapan, Indonesia
Author Emails
a)
Corresponding author: wibisono@universitasmulia.ac.id
b)
maw.wardhana@ universitasmulia.ac.id
c)
wisnuhera@ universitasmulia.ac.id
Abstract. Academic activities require supporting devices that are always ready, both hardware and software, and networks.
However, the device may be experiencing problems. Fast problem solving is needed so that academic activities continue to run
well. The problem is that the asset data is not integrated with the helpdesk, the difficulty of prioritizing work, and there is no
previous troubleshoot recording as reference material. The Extreme Programming approach is used to make the development
process more efficient, with iterations carried out at each stage. The SWOT Analysis and Value Chain Analysis helps identify the
requirements correctly. Using Google Apps as a development tool for the Helpdesk Ticketing System proves to be more efficient
as there is no need to invest in server hardware for processing and storage. Google Apps Script is also reliable enough to
complete various processes designed in this system. Implementation of Google Apps is also easier because users are familiar with
Google Forms and Google Sheets interfaces. With a good helpdesk ticketing system process, the implementation of academic
activities will also run well.
Keyword: helpdesk, ticketing, extreme programming, google apps script, swot analysis, value chain analysis.
INTRODUCTION
The implementation of academic activities requires the support of software, hardware, and computer networks.
Software is academic information systems; hardware is personal computers, laptops, printers, and other supporting
hardware. Computer networks include networks between computers and access to the internet in wired or wireless
access.
In daily operation, each supporting device can experience problems. Any problems that arise need to be addressed
immediately so that academic service activities run well. However, several problems can arise simultaneously, and it
is not easy to solve them at one time. Problems that arise may have happened before, and technicians rely on their
memory of solutions that have been done because there is no good documentation. From the previous reseach 1, it is
known that users also need to know the progress of the completion of the report. The problem of this research is that
the device asset data has not been integrated with the helpdesk section, there is no guideline for determining the
priority of work, and there is no troubleshoot documentation that has been done.
070005-1
Literature Review
Many helpdesk applications have been developed. The helpdesk applications for Albanian radio and television
using Java2, designing a helpdesk ticketing system for retail companies3, designing a helpdesk ticketing system using
the waterfall method for property companies4, and implementing a helpdesk ticketing system for software companies
using the approach ITIL5.
SWOT Analysis can be used for strategic planning and can be integrated with other methods6, has also been applied
in various studies such as application of mass spectrometry in clinical laboratories7, and determine appropriate
strategies for implementing prefabrication in rural areas8.
Value Chain Analysis is widely used in previous studies to calculate financial effects on food chain actors to
improve animal health and welfare9, examine the value chain of liquid milk10, and identify and understand the
weaknesses of the seaweed policy11. The combination of SWOT Analysis and Value Chain Analysis was carried out
in research12 to design a fire spot detection system.
Extreme Programming is an application development method that is a highly time-efficient and lightweight
software development technique based on principles of ingenuousness, interaction, feedback and strength 13. It is also
very responsive, fast, and suitable for developing a system with requirements that often changes to user needs 14.
Although there are not many, several previous studies on Google Apps implementation have been carried out.
Previous research includes implementing Google Suite Apps to monitor student engagement in a complex delivery
situation15, implementing controlled digital lending with Google Drive and Apps Script16, and the implementation of
Google Apps to enhance the achievement of Online Collaborative Learning for students 17.
METHOD
This study uses a qualitative method. Data will collect through observation, interviews, and literature studies.
Observations on the activities carried out in the troubleshooting process, while interviews will conduct on users,
helpdesk, and technicians. The literature study will conduct by studying the relevant journals. After the data and
information are obtained, a SWOT analysis will be carried out to see the strengths, weaknesses, opportunities, and
threats so that strategies can be prepared to improve the capabilities of existing systems and become the basis for
developing new systems. The new system's design begins by creating a Value Chain Analysis model to describe the
purpose of the system and the main and supporting activities. Furthermore, based on the Value Chain Analysis model,
a new system was developed using the Extreme Programming method, which emphasized developing software
efficiently through practical principles and techniques. The research model is shown in Figure 1.
Figure 1. Research Model
Extreme Programming is one of the methods included in Agile Methodology, emphasizing efficiency and process
speed, providing quality results and convenience for programmers13 and providing solutions when users have not been
able to properly define their needs and allow for changes in the development process14.
Based on the description of Extreme Programming in articles18 and articles19, original models, and modified
models, the Extreme Programming model was simplified in Figure 2.
070005-2
Figure 2. Extreme Programming Model
There are four main stages in the Extreme Programming method: Planning, Design, Coding, and Testing. Planning
is developed based on the SWOT analysis results, including problem identification, needs analysis, and system
development schedule. At the Design stage, a new system modelling is carried out based on the Value Chain analysis
made, consisting of system modelling, architectural modelling, and database modelling. System and architecture
modelling uses the Unified Modeling Language (UML) while database modelling At the Coding stage, modelling is
applied to the user interface using Google Apps. While the Testing stage, testing the program's functionality using the
Blackbox testing method.
Based on the user requirements defined at the beginning, planning, design, coding, and testing will be carried out,
each of which can be repeated or iterated. In the testing process, when the user has agreed, the section will be
immediately released without waiting to complete the entire process.
FINDINGS AND DISCUSSION
Data Collecting
Data are collected by interviewing all parties involved in the troubleshooting process, including the Helpdesk,
technicians, and users. The focus of the interview is to identify activities that have been carried out in handling user
problems. Users explain the often experienced problems, how to report to the helpdesk, and the level of satisfaction
with the services received. Next, dig up information from the Helpdesk Officer about how the process receives
problem reports from users and forwards them to technicians to work on solutions. For technicians, interviews focus
on the problem-solving process, the most frequent problem, and how users respond to the actions taken.
Direct observation of the process in the field was carried out to complete the information. Observations were made
on the user reporting process, the process of receiving reports at the helpdesk, and the repair process at the technician.
This direct observation is essential to know more about the current process. From the results of interviews and
observations, it can be concluded that, in general, handling user problems is still ongoing manually, and there is no
information system used.
SWOT Analysis
The next step is to conduct a SWOT analysis to see the strengths, weaknesses, opportunities, and threats in
implementing the existing helpdesk services. The results of the SWOT analysis are shown in Table 1.
070005-3
Table 1. SWOT Analysis
STRENGTHS WEAKNESSES
1. The university has sufficient IT 1. The number of technicians is still lacking.
equipment assets in providing 2. Does not have a hardware asset database.
academic services. 3. Does not have an information system for the
2. The skill of the technician team to helpdesk.
troubleshoot is quite good. 4. Does not have a database of previous
troubleshoot.
OPPORTUNITIES THREATS
1. The development of technology and 1. Technological developments will have an impact
the internet makes it easier to do on increasing the number of devices that
troubleshooting. institutions will own.
2. The emergence of practical tools such 2. Competitors will also develop their academic
as Google Apps that facilitate services supported by the availability of an
programming. effective helpdesk system.
Value Chain Analysis
A Value Chain Analysis model is used to determine the business process of the helpdesk ticketing system to be
built, which describes the main activities in the business process and supporting activities. In this helpdesk ticketing
system, the main activities are ticket opening, technician assignment that will be carried out by the helpdesk officer,
problem-solving process, and user assessment related to solving that is carried out. Supporting activities consist of
Asset Procurement, Data Asset Management, User Management, and Notification System. The Value Chain Analysis
model for the helpdesk ticketing system can be seen in Figure 3.
Figure 3. Value Chain Analysis
From Figure 3, it can be seen that the main activities and supporting activities in the designed helpdesk ticketing
system. The main goal of the entire series of business processes is to provide the best service for users. This Value
Chain Analysis will continue with system development using Extreme Programming with stages: Planning, Design,
Coding, and Testing, each of which allows for iteration according to user needs. In Extreme Programming, it is
possible to make minor releases of modules that the user has approved without waiting for the entire module to finish.
Planning
The first step in the Extreme Programming method is planning based on the explanation of user needs. At this
stage, a problem identification study will be conducted based on interviews and observations of the helpdesk ticketing
system and SWOT analysis. The need for a new ticketing helpdesk system that has been defined in the Value Chain
Analysis will be further developed, and determine the system development schedule.
070005-4
Design
At this stage, a design is made for application development using Google Apps. The first step is to design the
database, which will be a table in Google Sheets. The database model is shown in the Conceptual Entity Relationship
Diagram, as in Figure 4.
Figure 4. Conceptual Entity Relationship Diagram
From Figure 4, it can be seen that the database design required is Employee, Device, Ticket, Problem, and
SolvingBook. Employee data directly use employee data owned by the university, while Device data uses data owned
by the Asset department, especially hardware supporting academic services. Ticket data accommodates every
complaint ticket made by the user, supported by the Problem table, which contains the classification of problems. If
the incoming ticket contains a new problem or a new solution to an existing problem, the Technician can record it
specifically and accommodate it in the SolvingBook table.
Furthermore, it describes how interactions occur in the designed helpdesk ticketing system, using the Use Case
Diagram approach as shown in Figure 5.
Figure 5. Use Case Diagram
The activities carried out in the system are depicted in an Activity Diagram as shown in Figure 6.
070005-5
Figure 6. Activity Diagram
The activity begins with the User opening a ticket which the Helpdesk Officer will directly read. After confirming
the type of problem encountered, the Helpdesk Officer will assign it to the appropriate Technician. The Technician
will troubleshoot the tickets received and report when the work has been done to the User. Based on the report received,
the User will provide an assessment based on two aspects, namely response time and solving quality. Meanwhile, the
Asset Officer will perform asset data maintenance related to new device data, asset transfers, and asset write-offs.
Coding & Testing
With the Extreme Programming methodology, coding and testing can run more flexibly. Users can do testing at
any time every single module has been completed without waiting for all modules. Each stage of the process can
iterate as needed.
Programming in this system completely uses Google Apps. Google Forms with Google Apps Script support is
used for the ticket opening process, assignments by the Helpdesk Officer, filling out Solving Reports by technicians,
and Maintenance Data Assets by the Asset Officer. Database and programming are done in Google Sheets which
Google Apps Script supports. The role of each Google Apps application in the helpdesk ticketing system is shown in
Figure 7.
Figure 7. Google Apps Implementation
The testing process for this system uses Blackbox testing, which focuses on testing program functionality. With
the Extreme Programming methodology, every module that has been tested and approved by the user can be used
immediately without having to wait for the entire module.
070005-6
CONCLUSION
Although Google Apps is not as comprehensive as programming languages, it effectively solves simple needs.
Implementation of Google Apps will also reduce investment for universities because server hardware is not required.
This study shows the effective implementation of Google Apps in constructing an efficient helpdesk ticketing system.
Problems identified, namely the non-integration of asset data with the helpdesk, work priorities, and previous
troubleshoot references, can be solved by implementing Google Apps. Making SolvingBook tables with Google
Sheets is very efficient because technicians can quickly access previous troubleshoots with similar cases.
REFERENCES
1
A.S. Dzihni, R. Andreswari, and M.A. Hasibuan, Procedia Comput. Sci. 161, 903 (2019).
2
F. Ismaili, A. Bulku, and B.A. Caushi, 2018 41st Int. Conv. Inf. Commun. Technol. Electron. Microelectron.
MIPRO 2018 - Proc. 485 (2018).
3
J.C. Wibawa, E. Prasetyo, and R. Fauzan, IOP Conf. Ser. Mater. Sci. Eng. 662, (2019).
4
S.N. Khasanah, S.J. Kuryanti, Hermanto, and C. Adiwihardja, J. Mantik 4, 56 (2020).
5
A.S. Girsang, Y. Kuncoro, M.H. Saragih, and A.N. Fajar, IOP Conf. Ser. Mater. Sci. Eng. 420, 0 (2018).
6
M.A. Benzaghta, A. Elwalda, M. Mousa, I. Erkan, and M. Rahman, J. Glob. Bus. Insights 6, 55 (2021).
7
M. Vogeser and Y.V. Zhang, Clin. Mass Spectrom. 9, 1 (2018).
8
J. Zhou, P. He, Y. Qin, and D. Ren, Sustain. Cities Soc. 50, 101715 (2019).
9
J. Niemi, R. Bennett, B. Clark, L. Frewer, P. Jones, T. Rimmler, and R. Tranter, PLoS One 15, 1 (2020).
10
A. Thakur, A.K. Dixit, S. Kumar, and G. Bhandari, Agric. Res. 10, 307 (2021).
11
A.M. Nor, T.S. Gray, G.S. Caldwell, and S.M. Stead, J. Appl. Phycol. 32, 2161 (2020).
12
Y. Wibisono, EDPACS EDP Audit. Control. Secur. Newsl. 64, 1 (2021).
13
A. Shrivastava, I. Jaggi, N. Katoch, D. Gupta, and S. Gupta, J. Phys. Conf. Ser. 1969, (2021).
14
B.G. Sudarsono, S.P. Lestari, A.U. Bani, J. Chandra, and J.F. Andry, Int. J. Emerg. Trends Eng. Res. 8, 2223
(2020).
15
Anthony Ward, in 2021 30th Annu. Conf. Eur. Assoc. Educ. Electr. Inf. Eng. (IEEE, Prague, Czech Republic,
2021).
16
X.W. Qinghua Xu, Leon Lin, Int. J. Librariansh. 6, 37 (2021).
17
N. Tongchai, in Proc. 13th NPRU Natl. Acad. Conf. (The 13th NPRU National Academic Conference Nakhon
Pathom Rajabhat University, Nakhon Pathom, 2021), pp. 906–913.
18
L. Bossavit, Agil. Alliance 1 (2012).
19
J. Johnson, BMC Softw. 1 (2020).
070005-7
Application of System Usability Scale (SUS) Method in
Testing the Usefulness of Information System Student
Creativity Program (PKM) Based on Website
Silvia Oktaviani1,a) , Citra Wiguna1,b), Agus Priyanto2,c)
Author Affiliations
1
Information System Study Program, Informatics Faculty, Telkom Institute of Technology Purwokerto,
Indonesia
2
Informatics Study Program, Informatics Faculty, Telkom Institute of Technology Purwokerto,
Indonesia
Author Emails
a)
Corresponding author: 17103103@ittelkom-pwt.ac.id
b)
citra@ittelkom-pwt.ac.id
c)
agus_priyanto@ittelkom-pwt.ac.id
Abstract. One of the significant indicators in the Education field is having a good experience from outside the
campus. The student experience is reflected in three main factors: thoughts, feelings, and skills. Students Creativity
Program (PKM) is a form of the implementation of the Threefold Mission of Higher Education (Tri Dharma Perguruan
Tinggi). PKM has an effort to grow, accommodate, and realize students' creative and innovative ideas. The use of
information technology that is said to be a supporting factor in running business processes can improve effective and
efficient performance, one of which is the use of computerized websites in PKM activities at Telkom Institute of
Technology Purwokerto. The use of a website that is new still needs a lot of development. There are still some functions
that have not run as they should be, such as the assessment process on the website that is still semi-manual. Based on
these problems, so that the website can run optimally, it is necessary to evaluate the feasibility of the PKM system
used by end users at Telkom Institute of Technology Purwokerto. The test was conducted using the System Usability
Scale (SUS) method consisting of 10 statement items in the form of questionnaires with the main focus on measuring
the feasibility level of the website involving respondents as end-users. The results showed that the validity test is said
to be valid. Meanwhile, the reliability test is said to be reliable with a score of 63,308. This result can be categorized
OK with grade scale D. The result shows that the research can be developed by updating the content or interface of
the system to be more consistent and conducting demonstrations or socialization on how to use the system before users
access the system.
1. INTRODUCTION
Students Creativity Program (PKM) is a form of the implementation of the Threefold Mission of Higher
Education launched by the Directorate General of Higher Education under the management of the Directorate of
Learning and Student Affairs (Belmawa). PKM was launched as a form of effort to grow, accommodate, and
realize students' creative and innovative ideas. The new policy of the Minister of Education contained in
Kepmendikbud No. 754/P/2020 concerning 8 Key Performance Indicators (IKU) became the foundation of the
transformation of Higher Education in Indonesia. These indicators have an assessment based on several things
and those that are directly related to students are university graduates getting decent jobs and students getting
outside the campus experience. It is in line with PKM, where creativity is the integrative incarnation of three main
factors in human beings, namely thoughts, feelings, and skills. The mind factor includes imagination, perception,
and reason. The feeling factor consists of emotion, aesthetics, and harmonization, while the skill factor contains
080001-1
talent, body physiology, and experience. Students can reach the creative level if these three factors are optimally
pursued in the PKM activities that have been organized[1].
The utilization of information technology helps the education field in running business processes to support
decision-making on the problems faced and improve effective and efficient performance[2]. One form of
utilization of information technology in the form of a set of pages that display information in the form of text,
images, animation, sound, video, or a combination of these by forming a series of interrelated components and
connected through a network of pages is a website[3]. The use of a computerized website at the Telkom Institute
of Technology Purwokerto can meet the needs of users in carrying out PKM business processes. Therefore, they
run and are well documented to assist managers in processing data and make it easier for all parties to access
information and data related to PKM.
The use of the website that is still in the trial phase or a month running still needs a lot of development. This
is because there are still some functions that have not been running as they should be. One of them is related to
the assessment process on the website that is still semi-manual. Therefore it requires the assessor to download the
assessment form and re-upload the results of the assessment. Based on these problems, it is necessary to evaluate
and test its' usefulness to make the website run optimally so that it can be used easily by the users. This test was
conducted using the System Usability Scale (SUS) method. The SUS method consists of 10 statement items in
the form of questionnaires. This method provides convenience in the process, does not require complex
calculations, and is free of charge. The result is proven to be valid and reliable even though the sample size used
is small[4].
There are some previous researches in various fields using the SUS method. Research in the field of education
applies the SUS testing method. The result is that the average value of respondents is 76.25 with details of 0% of
respondents assessing that the application system is not accepted, 10% of respondents giving a marginal
assessment, and 90% of respondents stating the system can be accepted[5]. Further research has been applied in
the field of transportation with the SUS method. As a result, an average score of 79.5 is obtained, and the system
is declared acceptable to users[6]. Based on the research that has been put forward the SUS testing method can be
applied to various systems. The SUS method has a primary focus on measuring the feasibility level of the system
by involving respondents who are end-users.
.
2. METHODS
This study uses the descriptive method. Where a method that raises problems by collecting data is presented
to describe the characteristics of a situation or object of study and draw a conclusion[7]. The first stage is problem
observation. This stage is carried out by conducting a direct review of the object of study to determine the extent
to which the PKM business process has been running by conducting an outline analysis of the existing problems.
The second stage is data analysis. This stage is carried out by collecting data starting from the PKM
recapitulation data for the last 5 years as many as 473 registrants, data on recapitulation of proposals that passed
funding, PKM manuals, and related data to get more accurate results. This step is carried out to determine the
extent of the scope of the implemented website used.
The third stage is the stage of questionnaire dissemination. This stage is carried out by applying 10 statement
items that exist in the SUS method with a test scale starting from 1 to 5, where 1 means strongly disagree and 5
means strongly agree with the statement. The 10 items of statement as shown in Table 1 below.
Table 1 The SUS Statement Items [8]

Code Statement Items
R1 I will often use/visit this site
R2 I find this site too complex (contains a lot of unnecessary stuff)
R3 I find this site easy to browse
R4 I need technical assistance to use/browse this site
R5 I rate the functions/features provided on this site as well designed and
prepared
R6 I rate too many inconsistencies on this site
R7 I feel most people will easily use/browse this site quickly
R8 I find this site very complicated to explore
R9 I feel very confident browsing this site
R10 I need to learn a lot before I can explore this site properly
080001-2
The fourth stage is data processing. This stage is carried out by tapping the final result of the questionnaire to
find the average value of the SUS score where the statement is odd-numbered then the respondent's answer scale
is reduced by 1 and the statement is even-numbered then 5 minus the scale of the respondent's answer, after that
it is summed and multiplied by 2.5 to determine the average value[9]. Then conducted a data validity test using
Pearson to determine the validity/invalidity of the website used, by testing the correlation of the linear relationship
between the two data sets with the following formula[10].
݊σ‫ݕݔ‬െሺσ‫ݔ‬ሻሺσ‫ݕ‬ሻ
‫ ݕݔݎ‬ൌ (1)
ξሺ݊σ‫ʹݔ‬െሺσ‫ݔ‬ሻʹሻെሺ݊σ‫ʹݕ‬െሺσ‫ݕ‬ሻʹሻ
Description:
rxy: correlation coefficient (rcount)
n: number of subjects
x: score each item
y: total score
ሺσሻʹ: the square of the number of item scores
ሺσሻʹ: square of total score
σʹ: sum of squares of item scores
σʹ: sum of squares total score
And also conducted the reliability test using Cronbach's Alpha to find out the extent to which the website is
said to be reliable with the standard > 0.7[11]. This test can be interpreted as a positive relationship between the
questions with one another with the following formula[10].
݇ σৡड़ʹ
‫ ݎ‬ൌሾ ሿሾͳെ ሿ (2)
݅‫ݐ‬ ݇െͳ σৡੀʹ
Description:
rit: reliability coefficient
k: number of questions
σʹ: number of item variants
σʹ: total variance
The fifth stage is the analysis of the SUS score results. This stage is the final stage in the study conducted by
measuring the level of assessment results that have been carried out before by classifying them into several
categories. They are acceptability, grade scale, and adjective rating.
3. RESULT AND DISCUSSION
The dissemination of questionnaires has been conducted through social media to students, reviewers, and
admins. The questionnaire was performed online using the google form. The samples used in this test are 34
people with the number of admins 1, reviewers 2, and students 31, calculated from the average number of
applicants in the last year. The results of the questionnaire are calculated by the following formula:
ൌሺሺܴͳെͳሻ൅ሺͷെܴʹሻ൅ሺܴ͵െͳሻ൅ሺͷെܴͶሻ൅ሺܴͷെͳሻ൅ሺͷെܴ͸ሻ൅ሺܴ͹െͳሻ൅
ሺͷെܴͺሻ൅ሺܴͻെͳሻ൅ሺͷെܴͳͲሻൈʹǡͷሻ
The results show an average score of 63.308, as shown in Table 2 below.
Table 2 The SUS Score Calculation Results
Respondent R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 Score
1 3 2 4 4 3 3 4 2 4 4 57,5
2 4 2 4 4 4 2 4 2 4 4 65
3 3 2 4 2 3 2 3 2 3 3 62,5
4 3 3 4 2 4 2 4 2 4 3 67,5
5 3 3 3 3 3 3 3 3 3 3 50
6 3 4 5 3 4 4 4 1 4 4 60
7 3 3 4 2 4 3 4 3 5 3 65
8 3 3 4 4 4 4 5 2 4 4 57,5
080001-3
Respondent R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 Score
9 2 2 4 2 2 4 3 2 2 2 52,5
10 2 1 4 4 5 2 4 1 5 4 70
11 3 2 5 1 3 3 4 2 3 3 67,5
12 5 2 5 1 3 3 5 1 5 1 87,5
13 5 5 5 3 5 4 5 2 5 2 72,5
14 3 3 3 4 4 3 3 4 4 3 42,5
15 4 2 4 2 2 2 4 2 4 2 70
16 4 2 4 4 4 3 4 2 4 2 67,5
17 2 2 5 2 5 2 5 2 4 3 75
18 4 2 5 4 4 2 4 2 5 3 72,5
19 3 2 4 2 4 2 4 2 4 3 70
20 4 4 4 4 4 4 4 4 4 4 50
21 3 3 3 2 3 3 4 2 3 2 60
22 5 4 5 5 5 4 5 5 5 5 55
23 3 3 3 3 3 3 3 3 3 3 50
24 4 4 4 3 3 4 4 3 4 4 52,5
25 3 2 4 2 4 2 4 2 4 3 70
26 3 4 4 4 3 4 3 3 3 3 45
27 4 2 4 3 5 1 4 1 4 3 77,5
28 4 3 5 2 4 3 5 1 5 2 80
29 3 1 4 2 3 1 4 1 5 2 80
30 4 2 4 3 4 2 5 2 5 5 70
31 3 3 3 4 3 3 3 3 3 3 47,5
32 4 2 3 2 4 3 4 2 4 2 70
33 3 3 4 2 3 2 3 3 3 3 57,5
34 4 3 4 3 4 4 4 4 4 4 55
Average SUS Score 63,3088235
The second phase of the validity test was conducted manually using Pearson at a signification level of 5%. By
comparing Rcount with Rtabel, where if the number of Rcount > Rtable the result is considered valid, with an
Rtable of 0.339. Validity test results are shown in Table 3 below.
Table 3 Validity Test Results

Rcount Rtable Description
R1 0,56612 0,339 Valid
R2 0,64 0,339 Valid
R3 0,341 0,339 Valid
R4 0,653 0,339 Valid
R5 0,597 0,339 Valid
R6 0,523 0,339 Valid
R7 0,489 0,339 Valid
R8 0,501 0,339 Valid
R9 0,517 0,339 Valid
R10 0,573 0,339 Valid
Table 3 shows that Rcount on the total of 10 question items on the questionnaire was declared greater than
Rtable. Therefore, the 10 items of the statement were said to be valid. The reliability test using Cronbach's Alpha
is considered reliable if the resulting value is greater than 0.7. The reliability test is shown in Table 4 below.
080001-4
Table 4 Reliability Test Results
Cronbach’s Alpha Number of Items Description
0,754 10 Reliable
The results showed that the value of Cronbach's Alpha for the 10 items of the questionnaire statement was
0.754, this can be said to be reliable because the results are >0.7. Therefore, the questionnaire is considered
reliable. The average respondent's assessment result was 63,308 with the analysis of the final SUS score as shown
in Figure 1 below.
Figure 1. the SUS Score [12]
Figure 1 explains the determination of assessment results based on Bangor et al (2009) below.
1. Acceptability
Acceptability range results belong to the marginal high category with a score of 63.308 which means the
result can be accepted with a chance of reaching a high average as shown in figure 1[13].
2. Grade scale
Grade scale results belong to the grade D category with a total score of 63,308 which indicates the value
is still below the average.
3. Adjective rating
The result of the SUS score obtained from this study is 63,308 thus the adjective range in the category
OK which means the midpoint value between neutral and average, indicates that the system is
acceptable[13].
The condition can be caused by differences in viewpoints in the assessment process that has been carried out.
However, the results of respondents' assessment of the PKM website can be said to be good and the data is utilized
by the user. This result is shown at the time of the previous validity and reliability test.
CONCLUSION
Students Creativity Program (PKM) website test results with the SUS method of 63,308 show that the validity
test is valid because the Rcount value > Rtable. Meanwhile the reliability test is reliable because Cronbach's Alpha
value is 0.754 > 0.7. The results are still not fully usable and in the category of OK with grade scale D, it shows
that this study can be developed by updating the content or interface of the system to be more consistent, and
conducting demonstrations or socialization on how to use the system before users access the system.
ACKNOWLEDGMENTS
Gratitude is dedicated to (1) the Student Affairs of Telkom Institute of Technology Purwokerto as the object
of research, (2) supervisors who wisely provide guidance advice, criticism, and contributions in completing this
research, (3) the support of parents and friends who cannot be mentioned one by one and (4) related parties to this
study, therefore, this study can be completed properly.
080001-5
REFERENCES
[1] Kemendikbud, “Program Kreativitas Mahasiswa - Pedoman Umum,” Kewirausahaan (PKM-K), Vol. 3,
Pp. 89–117, 2021.
[2] Erma Tita, “Analisis Dan Perancangan Sistem Informasi Perpustakaan,” J. Sist. Inf., Vol. 8, No. 1, Pp.
966–977, 2016.
[3] J. Wahyu, A. Faraday, S. A. Wicaksono, And M. C. Saputra, “Pembangunan Sistem Informasi Manajemen
Akademik Sekolah Dasar ( Studi Pada SDN Wates Kabupaten Kediri ),” J. Pengemb. Teknol. Inf. Dan
Ilmu Komput. Univ. Brawijaya, Vol. 2, No. 11, Pp. 4777–4784, 2018.
[4] Supriyatna, “Penerapan Metode Extreme Programming Dalam Perancangan Aplikasi Perhitungan Kuota
Sks Mengajar Dosen,” J. Khatulistiwa Inform., Vol. 3, No. 1, Pp. 106–113, 2018.
[5] P. Studi, T. Informatika, F. T. Industri, And U. A. Dahlan, “RANCANG BANGUN APLIKASI
TRACKING BERKAS YUDISIUM Fiftin Noviyanto 2 , Rahmat Ardani 1 , 1,2,” No. 2013, Pp. 128–136,
2016.
[6] I. Dinda Prakoso, M. Sudarma, And I. M. Arsa Suyadnya, “Rancang Bangun Sistem Pelaporan
Pelanggaran Parkir Berbasis Android,” J. SPEKTRUM, Vol. 6, No. 2, P. 130, 2019, Doi:
10.24843/Spektrum.2019.V06.I02.P18.
[7] Rasmila, “Evaluasi Website Dengan Menggunakan System Usability Scale (SUS) Pada Perguruan Tinggi
Swasta Di Palembang,” JUSIFO J. Sist. Inf., Vol. 02, No. Vol 2 No 1 (2018): JUSIFO, Pp. 108–121,
2018.
[8] J. Brooke, “A Quick An Dirty Usability Scale,” Iron Steel Technol., Vol. 15, No. 8, Pp. 41–47, 2018, Doi:
10.5948/Upo9781614440260.011.
[9] J. Brooke, “SUS: A Restrospective,” J. Usability Stud., Vol. 8, No. 2, Pp. 29–40, 2013.
[10] I. Fanani, S. P. Djati, And K. Silvanita, “Pengaruh Kepuasan Kerja Dan Komitmen Organisasi Terhadap
Organizational Citizenship Behavior (OCB) (Studi Kasus RSU UKI),” Indones. Christ. Univ., Vol. 1, No.
1, Pp. 40–53, 2016.
[11] K. S. Taber, “The Use Of Cronbach’s Alpha When Developing And Reporting Research Instruments In
Science Education,” Res. Sci. Educ., Vol. 48, No. 6, Pp. 1273–1296, 2018, Doi: 10.1007/S11165-016-
9602-2.
[12] A. Bangor, T. Staff, P. Kortum, J. Miller, And T. Staff, “Determining What Individual SUS Scores Mean:
Adding An Adjective Rating Scale,” J. Usability Stud., Vol. 4, No. 3, Pp. 114–123, 2009.
[13] P. Jeff Sauro, “5 Ways To Interpret A Sus,” Measuringu. 2018.
080001-6
Employee Monitoring, Productivity, Privacy and
Harassment: Auditing Technology Training and Adoption in
Public Administration
Shahid Anjum1, a) and Zahidah Ismail 1)
Author Affiliations
1
Universiti Teknologi Brunei, Tungku Link Road, Gadong
Brunei Darussalam, BE 1410
Author Emails
a)
Corresponding author: anjumsw@hotmail.com
a)
shah.jum@utb.edu.bn
Abstract. Organizational intentions may be focused on enhancing productivity and managing digital age business risks
through the adoption of employee monitoring technology, it may, however, be perceived as an unnecessary intrusion of the
privacy space of employees at the workplace which may develop a reaction towards non-adoption and / or intentional
resistance towards such automation. Auditing employee intentions before the initiation of any change in the technology
space in the organization is a vital strategy for any sort of change management. In this context, current study surveys the
public sector employees in the economy of Brunei Darussalam. The study has three research objectives which may be
described as the organizational dimensions of the application of monitoring technology at the organization regarding having
privacy policies and training opportunities for employees, to understand the intentions of employees for using monitoring
technology. A survey of public sector employees has been conducted in Brunei’s economy using paper based questionnaires
as well as using Google forms which has generated 149 responses ready for analysis using Cronbach’s Alpha, Correlation,
and Linear Regression methods. The study has found that employees are not very enthusiastic to use employee-monitoring
technology for the sake of increasing organizational productivity or profitability. However, as the technology can decrease
the unnecessary harassment at the workplace, therefore, this variable has reflected their intention to adopt to even employee-
monitoring technology if implemented at the workplace.
INTRODUCTION
Frauds are unethical behaviours by dishonest individuals to obtain benefits in unacceptable way and some fraud
is come from the employees of the organisation itself (Bierstaker et al., 2006) which is the main reason to implement
the employee monitoring technology by many organisations to monitor their employees. One of the most important
of several reasons, including preventing fraud, for monitoring at an organization is to increase the productivity of the
employees ( Ella et al., 2016) because realizing that the employer is watching their activity, the employees will be
behave well in online activities (Glater, 2010). While some suggest of retaining and auditing only the meta data of
employee activities, Ella et al. (2016) has mentioned about viewing of employees computer screens in real time.
Employee monitoring technology, however, may have double sided impact on an organization i.e. it may either
enhance productivity if implemented through employees consent or it may fire back. This is because employees may
show enthusiasm towards technology or changes if these do not violate basic human concerns and rights, but
employees may resist to any change involving technology and may avoid adaptation because of several reasons,
including on privacy violation grounds. Instead of increasing productivity of employees, actually reverse can also
happen (Samaranayake et al., 2012). As the relationships between motivations of economic agents and performance
of public sector, organizations are being widely discussed in the literature of empirical economics where employee
080002-1
monitoring technology (EMT) has been proposed as one of the popular solutions in order to improve public services
delivery, besides pay for performance contracts (Cordella et al., 2016).
Top employee monitoring practices have been described in Holt et al. (2017) which lists these as: access panels,
attendance and time monitoring, computer monitoring programs, email and text message monitoring, filters and
firewalls, GPS and RFIDs, internet and clickstream data monitoring, keystroke logging, SN network and search engine
monitoring and video surveillance. In our study, we are more interested to conduct a research on employee monitoring
software technologies. Some employee may show enthusiasm with change management efforts of the organizations,
but some will resist to changes as well as to adapt for the implementation of any new technology at the workplace for
several factors including on privacy violation grounds (Huth, 2013). It is important to gain employees’ trust because
employees’ behaviour is said to be affected by employees’ trust in the organisation (Brown, et al., 2015). Employees
are, however, having the feelings of becoming more vulnerable for trusting the employers as in recent era the
employers possess more power (Yue et al., 2019). Therefore, the employer should establish their trustworthiness in
order not to lose employees’ trust. Employees’ commitment can be increased if the employees trust the management
of the organisation (Constantin & Baias, 2015). However, employees nowadays have lack of trust towards employer
(Jiang & Probst, 2015). The public sector employees can be assumed to pose more resistance to such productivity
enhancing intentions by making use of employee monitoring technology than private sector because of the nature of
permanent employment contract in public sector. Therefore, it is essential to study and justify how employees adopt
to the new technology (Jan et al.2, 011) and this is the context which has provided the rationale for investigating such
issues among public sector employees of Brunei Darussalam as well. The study has tried to perform research on three
objectives pertaining to behavioural intention from the perspective of the public sector employees towards the
application of monitoring technology at the organization, monitoring technology usage, and the underlying logic and
factors influencing employee monitoring technology usage. Investigation of the variables like privacy policy, training,
productivity, profitability and harassment have been used along with conventional variables like perceived usefulness,
ease of use of technology and subjective norm which have been used in various versions of models of theory of
planned behaviour including in various versions of models of planned behaviour, reasoned action and technology
acceptance. The paper has been organized into five sections. After section 1 on introduction, section 2 reviews
literature and section 3 discuss the model and data. Description of results and discussion are presented in section 4.
Finally, section 5 concludes the study.
REVIEW OF LITERATURE
Employee Monitoring Technology is a system installed on employees’ computer to monitor the employees’
activities by collecting, storing, analysing and reporting the employees’ activities in the workplace (Samaranayake et
al., 2012). When an organisation decides to install the employee monitoring technology to the workplace, the
organisation should consider the legal and ethical issues that will be arise as the employee monitoring technology is
implemented (Yerby et al., 2013). There several activities that are considered as fraud such as theft of cash and bribery,
while the occupational fraud is where the culprit is the employee of the organization itself which normally consists of
three categories; asset misappropriation, corruption, and fraudulent financial statement (Bae et al., 2018).
Traditionally, an organisation will use auditing method to combat crime within organisation by monitoring the
employee manually, however, the method is time consuming (Bae et al., 2018). Therefore, employee monitoring
technology will be required to track down the culprit of fraud (Bierstaker, et al., 2006). Employee monitoring
technology, however, can caused negative feelings by the employees as the employees might think that their privacy
is invaded (Tomczak et al., 2017). Although the implementation has risen many issues by the employees, the
organisation has many reasons to use the technology in the workplace which includes preventing misuse of company
resources and preventing from breaching (Ariss, 2002). While customer services enhance the confidence of the
customers on the organization and thus enhance the profitability and on-going concern of the organization, monitoring
is also important in order to check the activities of the employees to prevent unethical behaviour from happening such
as sexual harassment and obtaining sensitive information for advantage (Moussa, 2015). Therefore, EMP is helpful in
reducing the workplace harassment as is estimated in Ismail et al. (2021). Three research objectives of the study are
described as the effect of organizations having policies and training opportunities for employees on the use of EMT.
The conventional TAM variables used in various models of theory of planned behaviour are perceived usefulness
(PU), perceived ease of use of technology (EOU) and subjective norm (SN). Theory of reasoned action has built on
the insights from the disciplines of psychology and social psychology. This theory has included various variables in
080002-2
order to explain the actual usage of technology like attitudinal belief, normative belief, attitude, subjective norm and
behavioural intention. A model of technology acceptance follows this, where user motivation, for actual system use,
is explained by the variables of perceived usefulness (PU), perceived ease of use (PEOU) and attitude towards using
(ATU), have been used. Subjective norm is represented by friends, family, colleague and peer categories and finally
perceived behaviour control is a product of resource facilitating conditions and self-efficacy. The study has tried to
perform research on the behavioural intentions from the perspective of the public sector employees towards the
adoption of monitoring technology if employed at the organization, towards its usage, and the underlying logic and
factors influencing such a usage.
DESCRIPTION OF MODEL AND DATA

This study has chosen a variant of technology acceptance model similar to Mugo et al., (2017) as our initial
proposed model. The model has been built based on the guidelines provided in broader literature of Technology
Acceptance Model in general and on the model proposed by Mugo et al. (2017) in particular. Mugo et al. (2017) has
believed that PU and PEOU can be modified or could be affected by certain internal and external factors that could
lead the consumers to have behavioural intention to use to the new technology. In our study, internal/external variables
part has been replaced with the understanding of the employees (UNDR) about the employee monitoring technology.
Besides conventional TAM variables, this study has also used privacy policy (PP), training of employees (T), and
organisational productivity, confidence of customers on organisation, customer confidence and harassment at
workplace as study dimensions. Organisation culture of policies section in the questionnaire serves the purpose of
studying whether the organisation has provided the employees a written privacy policy and training on handling
confidential information, culture will give impact to the acceptance of new technology (Sunny et al., 2019). Ease of
use section is to study whether the employees perceive that the employee monitoring technology is easy to use.
Perceive of usefulness section is to study whether the employees will perceive that the employee monitoring
technology will be useful for either to them, the organization or to both. Subjective norm section is to study that by
whom the employees are influenced by in regards the adoption of EMT. Understanding of EMT section is to study
the employees really understand what employee monitoring technology is and how it really works.Subjective norm
variable in this research model means trust of employees towards SN sections, which may influence their adoption of
EMT. It has three components i.e. family, department, and the government. The main reason why subjective norm
variable has three categories is to facilitate the study to whom the employees put more trust on. Therefore, this study
will see which group of SN can convince the employees that the employee monitoring technology can be helpful and
convenient. Intention to Use concept of the EMT has been represented through three of the proxies for intention to
use variables i.e. increase productivity of organisation, increased confidence of customers on organisation, and
decreased unnecessary harassment at workplace which are represented by ITU1, ITU2 and ITU3 respectively.
Fig. 1. Estimated Model
Therefore, the study has used a questionnaires based on Technology Adoption Model (TAM) which has been
distributed among government employees which has generated 170 responses out of which 149 sample size has been
used for analysis after rejecting incomplete responses. Various data analysis tools, such as Cronbach’s Alpha,
Correlation, Linear Regression and multiplicative regressions have been estimated in this study. The questionnaire
has been divided into six sections in which seven questions are in demographic section of the questionnaire, seven
questions representing organisational culture related to policies, four questions representing ease of use, six questions
for perceive of usefulness, nine questions in subjective norms section representing family, department and government
and ten questions representing the understanding of employees for monitoring technology. Each question from all
sections except for demographic section will be measured through 5-points Likert scale.
080002-3
Table 1. Various Regressions – Part 1
Regression Results for full data set of government sector (N=149)
IV A-R2 Beta t IV A-R2 Beta t
Regression 1: Dependent Variable = UNDR Regression 6; Dependent Variable = POU
PP 0.351 4.172* EOU 0.469* 0.692 7.110*
0.164*
T 0.120 1.428 Regression 7; Dependent Variable = ITU
Regression 2; Dependent Variable = UNDR POU 0.851* 0.905 27.183*
PP1 0.373 4.641* SN 0.056 1.696***
0.207* - Regression 8; Dependent Variable = ITU
PP2 -0.165
2.046**
Regression 3; Dependent Variable = EOU POU 0.852* 0.904 27.553*
PP 0.227* 0.482 6.669* SN3 0.067 2.048**
Regression 4; Dependent Variable = EOU Regression 9; Dependent Variable = UB
PP1 0.419 5.419* ITU 0.131* 0.370 4.828*
0.269*
PP2 -0.192 -2.487* Regression 10; Dependent Variable = UB
Regression 5; Dependent Variable = EOU
UNDR 0.223* 0.478 6.594* ITU3 0.153* 0.399 5.272*
Regression Results for data set of ML observations only (N=87)
Regression 1; Dependent Variable = UNDR Regression 6; Dependent Variable = POU
PP 0.329 3.058* EOU 0.558* 0.751 10.472*
0.135*
Regression 2; Dependent Variable = UNDR POU 0.875 19.624*
0.858*
PP1 0.361 3.366* SN 0.114 2.545*
0.190*
PP2 -0.167 -1.553 Regression 8; Dependent Variable = ITU
Regression 3; Dependent Variable = EOU POU 0.877 20.070*
0.859*
PP 0.234* 0.493 5.227* Government 0.119 2.715*
PP1 0.450 4.485* ITU 0.123* 0.365 3.617*
0.290*
PP2 -0.184 -1.832***
UNDR 0.297* 0.552 6.110* ITU3 0.161* 0.413 4.176*
Regression Results for data set of FL observations only (N=62)
PP 0.388 2.759* EOU 0.661* 0.817 10.960*
0.174*
Regression 2; Dependent Variable = UNDR POU 0.846* 0.921 18.364*
PP1 0.451 3.156* Regression 8; Dependent Variable = ITU
T1 0.210* -0.319 -1.477 POU 0.926 18.204*
T2 0.372 1.843*** Department 0.850* -0.269 -1.683***
Regression 3; Dependent Variable = EOU Government 0.288 1.813***
PP 0.200* 0.461 4.026* Regression 9; Dependent Variable = UB
Regression 4; Dependent Variable = EOU ITU 0.130* 0.380 3.181*
PP1 0.368 2.957* Regression 10; Dependent Variable = UB
0.218*
PP2 -0.210 -1.685*** ITU1 0.072** 0.296 2.401**
Regression 5; Dependent Variable = EOU ITU2 0.131* 0.381 3.194*
UNDR 0.127* 0.375 3.137* ITU3 0.132* 0.383 3.208*
Notes: Independent (Dependent) Variable = IV (DV); Adjusted R2 = A-R2; Beta Coefficient = Beta; t-Value (t-sig) = t;
EMPIRICAL RESULTS AND DISCUSSION
Linear regression has been used as an estimation technique after reviewing various options like stepwise linear
function technique (Anjum et al. 2021), logistic regression (Anjum et al 2020a and 2020b), advanced econometric
analysis (Shamim et al. 2015, 2013), regressioneering and regressgineering (Anjum 2014a and 2014b respectively),
080002-4
Table 2. Various Regressions – Part 2
Regression Results for data set of MS1 observations only (N=92)
Regression 1; Dependent Variable = UNDR Regression 7; Dependent Variable = ITU
PP 0.504 4.918* POU 0.902 22.955*
0.220* 0.875*
T -0.043 -0.420 SN 0.093 2.379**
PP1 0.216 0.474 5.108* POU 0.904 23.679*
Regression 3; Dependent Variable = EOU Department 0.881* -0.170 -1.664***
PP 0.255* 0.513 5.665* Government 0.269 2.639*
PP1 0.376 3.405* ITU 0.159* 0.410 4.264*
0.266*
PP2 -0.216 -1.953*** Regression 10; Dependent Variable = UB
Regression 5; Dependent Variable = EOU ITU1 -0.229 -1.485
0.229*
UNDR 0.233* 0.492 5.359* ITU3 0.660 4.289*
Regression 6; Dependent Variable = POU
EOU 0.657* 0.813 13.255*
Regression Results for data set of MS2 observations only (N=57)
PP 0.085 0.616 POU 0.780* 0.885 14.127*
0.172*
T 0.402 2.898* Regression 8; Dependent Variable = ITU
PP2 -0.205 -1.752*** Family -0.118 -0.511
0.265* 0.769*
T2 0.458 3.915* Department 0.165 0.572
Regression 3; Dependent Variable = EOU Government -0.065 -0.286
PP 0.176* 0.437 3.600* Regression 9; Dependent Variable = UB
Regression 4; Dependent Variable = EOU ITU 0.040*** 0.238 1.820***
PP1 0.476 4.078* Regression 10; Dependent Variable = UB
0.251*
PP2 -0.173 -1.485 ITU1 0.144 0.885
Regression 5; Dependent Variable = EOU ITU2 0.008 0.169 0.778
UNDR 0.188* 0.450 3.736* ITU3 -0.030 -0.129
Regression 6; Dependent Variable = POU
EOU 0.469* 0.692 7.110*
Regression Results for data set of IC dataset (N=149)
PP 0.351 4.172* EOU 0.602* 0.777 14.987*
0.164*
0.851*
PP1 0.373 4.641* SN 0.056 1.696***
0.207*
PP2 -0.165 -2.046** Regression 8; Dependent Variable = ITU
Regression 3; Dependent Variable = EOU POU 0.904 27.553*
0.852*
PP 0.227* 0.482 6.669* Government 0.067 2.049**
PP1 0.419 5.419* ITU 0.131* 0.370 4.828*
0.269*
PP2 -0.192 -2.487* Regression 10; Dependent Variable = UB
ITU3 0.153* 0.399 5.272*
Regression 5; Dependent Variable = EOU
UNDR 0.223* 0.478 6.594*
efficiency analysis (Shamim et al., 2017), data mining based logic (Anjum 2014c, 2013). Cronbach’s alpha results
have shown that maximum number is 0.989 and minimum is 0.793, thus highly significant. Regressions have been
estimated several times with different variables separately and simultaneously with different combination to see which
variables works well either individually or simultaneously. Therefore, after few iterations the empirical model shown
in Figure 1 and all regression results have been shown in Table 1 and Table 2. Final estimated model has shown that
080002-5
UEMT supported ITU indirectly that it has to pass through EOU and POU. UNDR has supported ITU indirectly
through the effects via EOU and POU. In the linear regression estimation of the proposed model, it has been found
that privacy and training aspects, which have initially been thought of as enabling factors affecting the UNDR, POU
and EOU, have however, not worked out. Besides, multiplicative effects have also been proven as insignificant.
CONCLUSION
Privacy policy and training increase employees’ understanding of technology. Ease of use, and not the
understanding itself, directly affects perceived usefulness. Subjective norm has been tested at extended level and has
included, apart from conventional family level instruments, some non-conventional options like the motivational
effect of organization’s department level senior management and country’s political leaders on the intention to use.
Surprisingly, only this latter category of subjective norm has shown statistical significance and thus has an impact on
the actual usage of employee monitoring technology. This result is logical as the country looks towards its Royalty
for even behavioural guidance. Based on the result of linear regression, it indicates that the employee in government
sector, specifically in Brunei Darussalam, believes that the implementation of employee monitoring technology can
help in decreasing unnecessary harassment at the workplace. This finding helps to provide an insight that the
employees might perceive the implementation of this technology infer increase in security, which indirectly makes
them feel safer to work at the workplace. Behavioural surveillance is important to create a sense of security for
employees. While there might be other possible reasons for the harassment, but the researcher believes that it arises
from the behaviour of the employees itself. Hence, this signifies the importance of creating a sense of security for the
employees through the implementation of this technology. Interestingly, the results clearly indicate that the employees
might perceive the employee monitoring technology will not increase the productivity of an organisation even though
the analysis of the findings shows that perceive of usefulness contributed the most to the result.
REFERENCES
1. J. L. Bierstaker, R. G. Brody and C. Pacini, Accountants’ perceptions regarding fraud detection and prevention
methods. Managerial Auditing Journal, 21(5), 520–535, (2006). https://doi.org/10.1108/02686900610667283
2. V. John Ella, J.D., C., & P.C., J. L., Employee Monitoring And Workplace Privacy Law American Bar
Association. Section of Labor and Employment Law National Symposium on Technology in Labor &
Employment Law, 14, (2016).
3. J. D. Glater, Employee Monitoring Research Resources. New York, 1–16, (2010).
4. V. Samaranayake and C. Gamage, Employee perception towards electronic monitoring at work place and its
impact on job satisfaction of software professionals in Sri Lanka. Telematics and Informatics, 29(2), (2012).
5. A. Cordella and C. Tito, Motivations, monitoring technologies, and pay for performance. Journal of Economic
Behavior and Organization (2016). http://dx.doi.org/10.1016/j.jebo.2016.10.016
6. M. Holt, B. Lang, and S. G. Sutton, Potential Employees’ Ethical Perceptions of Active Monitoring: The Dark
Side of Data Analytics. Journal of Information Systems, Vol. 31. No. 2. AAA. USA, Summer, (2017).
7. C. L. Huth, The insider threat and employee privacy : An overview of recent case law. Computer Law & Security
Review, 29(4), 368–381, (2013). https://doi.org/10.1016/j.clsr.2013.05.014
8. S. Brown, D. Gray, J. McHardy and K. Taylor, Employee trust and workplace performance. Journal of Economic
Behavior and Organization, 116, 361–378. ((2015). https://doi.org/10.1016/j.jebo.2015.05.001
9. C. A. Yue, L. R. Men, and M. A. Ferguson, Bridging transformational leadership, transparent communication,
and employee openness to change: The mediating role of trust. Public Relations Review, (Jan.), (2019).
10. A. U. Jan and V. Contreras, Technology acceptance model for the use of information technology in universities.
Computers in Human Behavior, 27(2), 845–851, ((2011). https://doi.org/10.1016/j.chb.2010.11.009
11. E. C. Constantin and C. C. Baias, Employee Voice - Key Factor in Internal Communication. Procedia - Social
and Behavioral Sciences, 191, 975–978, (2015). https://doi.org/10.1016/j.sbspro.2015.04.319
12. L. Jiang and T. M. Probst, Do your employees (collectively) trust you? The importance of trust climate beyond
individual trust. Scandinavian Journal of Management, 31(4), 526–535, (2015).
13. J. Yerby, M. Georgia and S. College, Legal and ethical issues of employee monitoring, 1(Feb.), 44-55, (2013).
14. D. L. Tomczak, L. A. Lanzo and H. Aguinis, Evidence-based recommendations for employee performance
monitoring. Business Horizons, (2017). https://doi.org/10.1016/j.bushor.2017.11.006
15. S. S. Ariss, Computer monitoring: benefits and pitfalls facing management, 39 (October 2000), 553–558, (2002).
080002-6
16. M. Moussa, Monitoring Employee Behavior Through the Use of Technology and Issues of Employee Privacy in America.
SAGE Open, 5(2), (2015). https://doi.org/10.1177/2158244015580168
17. D. N. Z. Ismail and S. Anjum. Privacy Policy, Training and Adaption of EMT to Curtail Workplace Harassment
in Organizations: An Application of TAM”, Proceedings of the Computational Intelligence in Information
Systems Conference (CIIS 2020), Advances in Intelligent Systems and Computing Series. Springer
18. S. Anjum, Statistical Software and Regression Diagnostic Reporting with Fuzzy-AHP Intelligent Zax", Lecture
Notes in Software Engineering, Vol. 2, No. 1, February, (2014a).
19. S. Anjum and Qaseem, N. Basel violations, volatility model variants and value at risk: Optimization of
performance deviations in banks. Economics and Business Letters 10(3), 240-248, (2021).
20. S. Anjum and Hanafi, E. W. B. A. M. Vulnerabilities and Digital Violations in Software Products: Logistic
Regression Analysis,IEEE Xplore, Proceedings of 2020 International Conference on Cyber Warfare and
Security (ICCWS). Institute of Electrical and Electronics Engineers. Denver. pp. 1-6, (2020a).
21. S. Anjum. DoS Attacks, Triad and Privacy: Software Exposures in Microsoft, Apple and Google," IEEE Xplore,
Proceedings of 2020 First International Conference of Smart Systems and Emerging Technologies
(SMARTTECH), Riyadh, Edited by Koubaa et al., IEEE Society. Washington. pp. 53-58, (2020b).
22. S. Anjum, Back testing VaR Violations, Be-ALAM Regression and Internal Models of Portfolio Variance
Forecasting,” Journal of Money, Investment and Banking, FRDN Inc., 29, Oct. (2014b).
23. S. Anjum, Composite Indicators for Data Mining: A New Framework for Assessment of Prediction Classifiers.
Journal of Economics, Business & Management, 2(1), (Feb. 2014c).
24. F. Shamim, N. Yamori and S. Anjum, Clicks Business of Deposit Taking Institutions: An Efficiency Analysis.
Journal of Economic Studies. 44(6). Emerald Publishing. Thomson Reuters, (2017).
25. F. Shamim, S. Anjum and A. B. Wakil, Banking Risk and Operating Efficiency Measures in the Era of IT.
Accounting and Finance Research. Sciedu Press. 4 (1), (2015).
26. F. Shamim and S. Anjum, Technology Diffusion in Japanese Finance Industry: An Exploration. International
Research Journal of Applied Finance, VI (12), (December 2013).
27. S. Anjum, Algorithms for Predictive Classification in Data Mining: A Comparison of Evaluation Methodologies.
Journal of Industrial and Intelligent Information (JIII), 1(2), (June 2013).
28. S. Sunny, L. Patrick and L. Rob, Impact of cultural values on technology acceptance and technology readiness.
International Journal of Hospitality Management, 77(March), 89–96, (2019).
29. J. Bae, H. Sub and M. Button, Exploring the impact of organizational investment on occupational fraud :
Mediating effects of ethical culture and monitoring control. Intl. Journal of Law Crime and Justice., (Oct. 2017).
30. D. Mugo, K. Njagi, B. Chemwei and J. Motanya, Technology Acceptance Model and its Application to the
Utilization of Mobile Learning Technologies. British Jour of Mathematics & Computer Science, 20(4), (2017).
080002-7
Improve Educational Marketing Strategy Through Use of
Digital Marketing Technology
Richki Hardi1,a), Suprijadi 1,b), Ririn Kusdyawati1,c), Agustinus Noertjahyana 2,d)
Author Affiliations
1
2
Petra Christian University, Surabaya, Indonesia
Author Emails
a)
b)
suprijadi@universitasmulia.ac.id
c)
ririn@universitasmulia.ac.id
d)
agust@petra.ac.id
Abstract. The general purpose of this research is to know the marketing management of information technology-based
education. This research method is descriptive qualitative with a direct experience approach. Data collection techniques
through in-depth interviews, observation, and document analysis. Data analysis techniques with data reduction steps, data
presentation, and concluding. The study results are expected to increase the marketing of information technology-based
education in universities; methods and marketing media based on information technology have attracted people to become
their children's choice in higher education. The educational leadership has maximized the involvement of teaching staff
and employees in information technology-based marketing. The supporting factors include the physical building,
curriculum, human resources, accreditation, achievements, extracurricular and varied student activities, unique advantages,
study tours abroad, marketing strategies, especially those based on information technology. Meanwhile, the inhibiting
factors are estimated to come from local government policies and competition from competitors.
INTRODUCTION
Educational marketing activities are one of the efforts to maintain the existence of educational institutions. This is
done in various ways, including exploring the potential uniqueness that can be raised or needed and, of course bringing
out the advantages of the school so that it is of interest to the community1.
Marketing is a leadership process in which individuals and groups realize the desires of organizational leaders
through strategies for creating offers and changing product value with others. It also applies to marketing management
in the world of education2.
One of the problems in the world of education, significantly higher education institutions, is the lack of students
from the targets set by higher education. Many universities from year to year experienced a decline in new student
admissions. One of the reasons is the management of the higher education institution, which pays less attention to the
marketing of educational services. Educational marketing includes appropriate methods, content and media to develop
educational marketing so that universities demand the public.
Departing from the description above, the researcher is interested in research to provide input to all higher
education institutions.
Marketing management for educational institutions is needed in line with the increasingly attractive competition
between universities. Marketing is necessary for educational institutions to build a positive image. If the institution or
university has a good impression in the eyes of the public, it will likely be easier to overcome competition3. So,
marketing is a process that madrasas must carry out to satisfy stakeholders and the community. The emphasis on
080003-1
providing satisfaction to stakeholders is something that every institution must do to compete. This marketing can be
seen from the various creative and innovative efforts of education providers to explore the uniqueness and advantages
of their schools so that they are increasingly needed and in demand by service users. Education. To attract prospective
students, a marketing strategy is required that sell educational services and how to approach them following the wishes
and satisfaction of consumers4. An institution that wants to be successful in the face of competition must continuously
practice marketing5.
Online marketing media using the basis of information technology among those used are:
Web site
Instagram
Facebook
You tube
Website
Whatsapp
Line
Twitter
Iklan di radio
Offline marketing media are also not left behind, including banners and banners, brochures and leaflets, student
activities, parent meetings and presentations to schools and colleges.
The impact of marketing through online technology networks such as websites, Facebook, Instagram and others is
positive. Among them can get students from far away.
LITERATURE REVIEW
Marketing strategy is a basic plan to achieve company goals6. Meanwhile, the KBBI is defined as a careful plan
of activities to achieve specific goals. Strategy is seen as a program that includes the goals to be completed,
accompanied by specific actions or steps to accomplish these goals to respond to the environment. Experts express
different meanings in their presentation and emphasis, but all of them have almost the same purpose. One of the
marketing that is often done in the marketplace, and the marketplace is part of e-commerce, where e-commerce is like
the centre of the marketplace so that the process of the marketplace is regulated by e-commerce. A marketplace is a
location for buying and selling products where sellers and consumers meet in one place. The seller will sell his goods
installs provided by e-commerce with the concept of a marketplace. The goods sold in the marketplace will then be
advertised by the e-commerce party to get potential consumers. The success of e-commerce with this marketplace
concept is determined by the large number of sellers and consumers who join the e-commerce website7.
The utilization of E-Commerce technology is a business mechanism that works electronically by focusing on
online business transactions and has the opportunity to build more humane and personalized relationships with
customers8. This fact can foster global competitiveness because the marketing system is no longer limited to certain
areas. So far, the marketing system has relied on direct interaction with consumers or customers and is still centred on
a particular location. This certainly makes it difficult for the management to increase sales and expand the scope of its
marketing area. The number of competitors is increasing, making it difficult for management to maintain its business
performance and achieve its profitability target level. There is difficulty accessing various information about products,
the latest price information, limited and inflexible promotional media, personalization barriers, and operational costs
that are increasing over time9.
The limited mobility and market liberalization force the management to innovate in expanding the target market
through the design and use of E-Commerce. Consumers or customers can make orders and purchases without
restrictions on place and time, responsive to current information 10.
Definition of Marketing Strategy is defined as a process of determining the plans of top leaders that focus on the
organization's long-term goals, accompanied by the preparation of a method or efforts to achieve these goals11.
One of the shortest definitions of marketing is meeting needs profitably. Broadly defined, marketing is a social
and managerial process by which individuals or organizations obtain what they need and want by creating and
exchanging value with others12.
Marketing strategy is defined as an analysis of strategy development and implementation of activities in the
strategy of determining the target market for products in each business unit, setting marketing objectives, and
developing, implementing, and managing marketing program strategies, determining market positioning designed to
meet the desires of target market consumers13.
080003-2
A marketing strategy combines the marketing mixes that entrepreneurs will apply to serve their market. This
marketing mix must be organized so that it will serve as an appropriate weapon in its comparison in the market against
competitors. Therefore, the gun must be adapted to market conditions and the conditions of competition it faces, where
the types of marketing mix combinations include; products, prices, promotions, and distribution channels14.
The marketing strategy is conceptually straightforward. It begins with an analysis of the market the organization
is considering. This requires a detailed analysis of the organization's capabilities, competitors' strengths and
weaknesses, economic and technological strengths affecting the market, and current and potential customers. So, the
marketing strategy is a way to achieve the company's or merchant's goals in marketing products to consumers.
Marketing strategy is also essential for the success of the business or business being run, regardless of the form of
trade or business, whether conventional or online, because marketing strategy is a plan to prepare how to run a
company or business15.
Marketing is a social and managerial process by which individuals and groups obtain what they need and want
through creating, offering and freely exchanging products and services of value with others. About marketing, strategy
is defined as a fundamental tool planned to achieve company goals by developing a sustainable competitive advantage
through the market entered and marketing programs used to serve the target market16.
Thus, a marketing strategy is a plan that is systematically formulated regarding marketing activities to serve as a
guideline for implementing marketing variables such as market identification, market segmentation, market
positioning and marketing mix elements17.
About education, services can be defined as the activities of educational institutions providing services or
delivering educational services to consumers in a way that satisfies them18. Marketing in educational services is a
social and managerial process of getting what is needed and wanted through the creation of offers the exchange of
products of value with other parties in education19. Marketing ethics in teaching offers quality intellectual services and
overall character building. This is because education is more complex, carried out with full responsibility; the results
of education refer to the future, fostering the lives of citizens the next generation of scientists in the future20.
METHODOLOGY
The research subject is the respondent or resource person who will be interviewed or observed. The research
subject's identity is the university's marketing staff, and one of the universities that are the place of observation is
Mulia University, Indonesia.
This study's data collection methods used three ways: marketing staff interviews, observation, and document
analysis. Meanwhile, this qualitative research data analysis technique uses three activities, namely data reduction, data
presentation and conclusion drawing.
.
DISCUSSION
For the success of an institution in the long term, the institution must create satisfying services such as the need
and want for its customers. To do this moving service, the agency establishes a marketing mix. The marketing mix is
the marketing elements that are interrelated, blended, organized and used appropriately so that the company can
achieve its marketing objectives effectively while satisfying the needs and desires of consumers.
The application of marketing in education, the focus of the application of marketing is how to bring services closer
to the wishes and satisfaction of students, which of course must be supported by the role of experts in their fields,
adequate resources and facilities, and continually improving the quality of graduates21.
Planning is the first step a manager must take. The planning function includes defining organizational goals,
developing overall strategies, and developing and coordinating activities to achieve expected goals. Planning in
educational marketing aims to reduce or compensate for uncertainty and upcoming changes, focus attention on goals,
ensure or get the process of achieving goals carried out efficiently and effectively, and facilitate control. The steps
that must be taken in the educational marketing planning process are as follows:
Identify the market or competitors, the first stage in educational marketing is identifying and analysing the market.
In this stage, it is necessary to conduct market research to determine market conditions and expectations, including
academic attributes that are of interest to educate consumers, including mapping from other schools and of the business
successes is determined by the ability to understand competitors. The output of these capabilities supports management
in deciding where to compete and how to position among competitors. The analysis is carried out by identifying the
080003-3
industry and its characteristics, identifying businesses within the industry, then evaluating each company, predicting
competitor activities, including identifying new competitors that might break through the market or market segments.
Competitive analysis is an attempt to identify threats, opportunities, or strategic problems that occur as a result of
changes in potential competition and competitors' strengths and weaknesses. Competition analysis is dynamic.
Competitive analysis is a continuous activity and requires the coordination of information. Businesses and business
units analyze competitors by using competitor intelligence systems.
Market Segmentation and Positioning Market segmentation divides a market into distinct buyers based on needs,
characteristics, or behaviour, who may require different products. While positioning is the characteristics and
differentiation of natural products, it is easier for consumers to distinguish service products from one institution to
another.
Determining the target market is an essential step in managing educational institutions. In a very diverse market,
it is necessary to decide what attributes are of primary interest to academic users. Calls can be classified based on
demographic, geographic, psychographic, and behavioural characteristics. Thus, it will be easier for schools to
determine market characteristics and needs marketing strategies. After knowing the character of the market, it will
determine which part of the market will be served.
Product differentiation, differentiation is an effective way to seek market attention. Of the many existing
educational institutions, parents will find it difficult to choose their child's school because the attributes of interest
between educational institutions are increasingly standard. Educational institutions should pressure other schools on
attractive packagings, such as logos and slogans. Internet facilities may be shared, but the guarantee of a safe and
clean internet will attract parents' attention. Making distinctions easily can also be done through forms of physical
appearance captured by the five senses that give a good impression, such as wearing an attractive uniform and a clean
school building. A differentiation strategy will position the organization uniquely to meet the specific needs of its
customers.
Conventional
Digital Learning
Learning Products
Products
Analysis of Digital
Learning Product
Business
Opportunities
Conventional Digital Learning

Learning Products Products
Conventional Digital Products

Market
Product Product
classification 1 classification 2
FIGURE 1. Marketing product classification
CONCLUSION
To improve the existence and image of an educational institution, education managers should be able to apply the
concept of a marketing strategy for educational services. Marketing educational services is a step to reform educational
institutions to market academic service products to the public by creating and offering service product innovations the
line with market expectations and demands. The purpose of marketing in education is to provide information to the
080003-4
public about the products of educational institutions, to increase public interest and interest in the effects of educational
institutions, to differentiate the outcomes of educational institutions from other educational institutions, to give more
assessments to the public with the products offered, and to stabilize the existence and the meaning of educational
institutions in society. In formulating a marketing strategy, you should pay attention to the expectations and needs of
students by listening to student complaints and looking for solutions that can improve existing marketing strategies.
With these activity steps, universities can achieve a balance in the operationalization of teaching in competing for the
market from many higher education providers.
ACKNOWLEDGMENTS
Thank you to all the academics of Mulia University who have fully supported this research to be completed. Thank
you also to all supporters and information providers, marketing parties and others who have helped provide a direct
experience so that researchers can fulfil the completeness of the data in this study. Hopefully, this research can be a
reference for future researchers.
REFERENCES
1. D. Oana, Stud. Bus. Econ. 14, (2019).
2. F. Ahmad MAE, A. Ghaffar Tahir, A.G. Tahir, S.A. A Rizvi, M.B. Khan, and F. Ahmad, J. Appl. Environ. Biol. Sci 7,
(2017).
3. N. Manea and M. PURCARU, Ann. Spiru Haret Univ. Econ. Ser. 17, (2017).
4. L. Gómez-Bayona, G. Moreno-López, and L. Machuca-Villegas, RISTI - Rev. Iber. Sist. e Tecnol. Inf. 2020, (2020).
5. C.A. Stachowski, Educ. Manag. Adm. Leadersh. 39, (2011).
6. N.F. Zuhaeda, Educ. Adm. Res. Rev. 3, (2019).
7. shopify, Ecommerce Definition - What Is Ecommerce (2019).
8. G. Nöldeke, (2020).
9. M. Mäki and T. Toivola, Technol. Innov. Manag. Rev. 11, (2021).
10. A. Molla and P.S. Licker, Inf. Manag. 42, (2005).
11. A. Ghandour, Int. J. Electron. Commer. Stud. 6, (2015).
12. B. Ola and I. Egho-Promise, Eur. J. Electr. Eng. Comput. Sci. 4, (2020).
13. M. Tsagkias, T.H. King, S. Kallumadi, V. Murdock, and M. de Rijke, ACM SIGIR Forum 54, (2020).
14. G. Liu, Alexandria Eng. J. 61, (2022).
15. Emarketer, Glob. Ecommerce Updat. 2021 - Insid. Intell. Trends, Forecast. Stat. (2021).
16. H.A. Al-Ababneh and S.A.S. Alrhaimi, TEM J. 9, (2020).
17. A.G. Tyurikov, A.N. Zubets, P. V. Razov, A.N. Amerslanova, and N. V. Savchenko, Humanit. Soc. Sci. Rev. 7, (2019).
18. D.K. Abdul-Rahman Al-Malah, H.H. Khudair Jinah, and H.T. Salim ALRikabi, Period. Eng. Nat. Sci. 8, (2020).
19. I. Shpolianskaya and A. Prokhorova, BRAIN-BROAD Res. Artif. Intell. Neurosci. 10, (2019).
20. A. V. Bogoviz, S. V. Lobova, A.N. Alekseev, V.N. Prokofiev, and I. V. Gimelshtein, Horiz. 27, (2019).
080003-5
Utilization of Social Media Application in Regency SMEs
Performance
1,2, a) 2, b) 2,c) 2,d)

Ivan Armawan , Heri Pratikto , Sopiah and Sudarmiatin
Author Affiliations
1
Department of Management, Economics and business, Mulia University, Balikpapan, Indonesia
2
Department of Management, Economics and business, Malang state University, Malang, Indonesia
Author Emails
a)
Corresponding author: ivan.armawan@universitasmulia.ac.id
b)
heri.pratikto.fe@um.ac.id
c)
sopiah.fe@um.ac.id
d)
sudarmiatin.fe@um.ac.id
Abstract. The decline in public buying interest has an impact on all business sectors in Indonesia, including MSMEs.
During the Covid19 pandemic, Regency MSMEs still use conventional systems, so new breakthroughs are needed to
increase turnover and unlimited marketing so that MSMEs can survive and advance. Digital marketing activities are
promotions and market searches through online digital media, one of which is social networking facilities. The purpose
of this research is to broaden the knowledge and skills of digital marketing, especially social media, in micro, small
and medium enterprises (MSMEs) to increase sales turnover. Participants are MSME Regency Balikpapan city; the
method used in this activity explains digital marketing materials, share experiences and discussions, and the
implementation of social media applications. This study indicates that the use of social media that is often used is
Facebook in displaying business profiles by 56.25% and WhatsApp (WA) by 50%, Instagram (IG) by 18.75% and
others by 25 % for product promotion. Based on the research results on the use of applications in pandemic that is ride
12,5%, Permanent 37,5% and doen 50%. MSME actors have not maximally used social networks, especially YouTube,
to promote their products.
INTRODUCTION
MSMEs or Micro, Small and Medium Enterprises are parties that carry out a business that is engaged in business
fields that touch the community's interests. MSMEs or Micro, Small and Medium Enterprises are currently considered
the most effective form of reducing poverty. Based on statistics and research conducted, MSMEs are the most
significant business actors. MSMEs have been legally regulated through Law Number 20 of 2008 concerning Micro,
Small and Medium Enterprises. MSMEs are the most important economic actors in the Indonesian economy and have
proven to be the driving force and safety of the national economy in times of crisis and the dynamics of economic
growth after the financial crisis.
Indonesia has experienced the condition of the Covid-19 pandemic for several months. The decline in public
buying interest has an impact on all business sectors in Indonesia, including MSMEs. MSMEs have a vital role in
building the economy, one of which is in Balikpapan. In the conditions of the Covid19 pandemic, many MSMEs are
still conventional in their marketing, so breakthroughs are needed to increase turnover and unlimited marketing so that
MSMEs can survive and advance. Digital Marketing is valuable for entrepreneurs to increase turnover when
conditions are adjusted to their business activities. The problems that occur are usually novice MSMEs do not
080004-1
understand how to start their business activities.
The Sepinggan Baru Regency Housing, South Balikpapan, is a residential area located in the middle of the
city of Balikpapan where most of the residents, especially housewives, have businesses that can be used as a support
for family income. However, they are still using conventional marketing methods in the current condition, so a plan
is needed online to be easily recognized by the general public.
Micro, Small and Medium Enterprises are the most significant economic actors in Indonesia. They have proven
to be a big part of securing the national economy in times of crisis and motivating economic growth after the financial
crisis. MSMEs can become the business sector with the most significant contribution to national development; MSMEs
always create job opportunities for domestic workers who are quite large, so that efforts to reduce unemployment can be
adequately resolved and greatly help the government.
Indonesia has been experiencing the Covid-19 pandemic for several months. The decline in people's buying interest
affects all business sectors in Indonesia, including MSMEs. MSMEs have a vital role in building the economy, one of
which is in Balikpapan. During this COVID-19 pandemic, many MSMEs whose marketing nature is still conventional,
so new breakthroughs are needed to increase turnover and unlimited marketing so that MSMEs can survive and move
forward. Digital Marketing is beneficial for entrepreneurs to increase turnover when conditions like this are adapted to
their business activities. The problem that occurs is that usually, novice MSMEs do not understand the right way to start
their business activities.
Digital Marketing
Human communication is changing a lot along with increasingly developing Digital Technology, including
acting and making decisions. Marketing activities have a significant effect on the development of Digital Technology.
The evolution of goods and services activities with the existence of digital-based marketing refers to the process of
finding, finding and building consumers by paying attention to every consumer behaviour that is currently developing
with the aim and hope of making it easier to promote brands keep consumers and maximize income.
The application of the digital marketing concept to MSME business actors is carried out to be able to achieve
the potential in producing high marketing between products and the target market to be completed.
Involvement in digital marketing is mandatory for Regency MSME actors' strategies to influence customers and build
brand awareness as pioneers of healthy competition between MSMEs. Some digital marketing concepts that MSMEs
mainly carry out include:
a. SEO
Search Engine Optimization or SEO is rapidly becoming a field of specialization as digital marketing
develops. And it's also the fastest-changing as search engines are constantly refining their algorithms to provide a better
search experience.
A significant part of SEO has content that is the best answer for your potential customers. Apart
from that, there are some essential elements that SEO content should have, such as:
1. Titles that contain at least 80 characters to avoid being truncated in search results pages.
2. A meta description that describes the content in 160 characters or less.
3. Use header text (H1 tags) for the most relevant text.
4. An alt description for the image that describes the content of the picture.
5. Enter descriptive text for the link (not just “Click Here” or “More”).
b. Optimizing content for website readers
It's essential to know how people read online versus print content. Online reading habits affect everything from
SEO to design.
People tend to 'scan' online content rather than read it from top to bottom. Long paragraphs and a lack of navigational
elements tend to make readers uncomfortable to linger on the website.
Keep paragraphs short and to the point. Just two or three short sentences will suffice. If you still want to explain the
section, try making it into bullet points.
Add H1 H2 H3 tags for each sub-topic or point to make it easier for readers to see the critical points.
c. Content planning
Studying the content planning process can help you achieve content mastery. You will write better content
because you fully understand the needs of your target audience and your business.
080004-2
d. Full funnel content strategy
There is the three-funnel model in marketing activities, namely Attract, Engage, and Convert. Buyers in the
Attract stage may not actively seek solutions. They may not even be aware that they have a problem.
Those in the Engagement stage are looking for solutions, researching their options. Those in the Conversion stage
have decided to buy and need a final push for an answer.
This model can be the basis for understanding the concept and concept of digital marketing, as we discussed earlier.
A digital marketer will develop content for each stage of the funnel, each with a different SEO strategy, amplification
strategy, and targeting.
We must define our funnels to map out the right set of strategies and messages.
e. Social media
Social media is the most important choice for many business people when they enter digital marketing. It can
be said that social media is an effective way to reach the target market by building a brand persona. Not only that, but
social media also makes it possible to find out what the target market likes and is looking for. Like SEO, marketing a
business organically on social media takes more time and effort, but it can provide much cheaper results in the long
run. To support accelerating the marketing process, currently, the use of social media advertising has also become
commonplace in digital marketing strategies. The cost is also not too high; you can create targeted ads.
f. Pay Per Click
Pay per click or PPC is a way to reach a wider audience and prioritize your content over others. Google, Bing,
and Yahoo all allow you to run pay per click.
g. Conversion Rate Optimization (CRO)
Conversion rate optimization (CRO) is the art and science of improving your online user experience.
Businesses often use CRO to get more conversions (leads, chats, calls, sales, etc.) from website traffic.
This research is based on the journal Social media adoption and its impact on firm performance: the case of the UAE
Syed Zamberi Ahmad Marketing & Management Department, Abu Dhabi University, Abu Dhabi, UAE, Abdul Rahim
Abu Bakar Marketing Department, College of Business Administration, Prince Sultan University, Riyadh, Saudi
Arabia, and Norita Ahmad Marketing and Information System, School of Business Administration, American
University of Sharjah, Sharjah, UAE.
LITERATURE REVIEW
The concept of digital marketing comes from the internet and search engines on websites. When internet usage
exploded in 2001, the market was dominated by Google and Yahoo, like search engine optimization (SEO). According
to the American Marketing Association (AMA), the definition of digital marketing is the activities, institutions, and
processes facilitated by digital technology to create, communicate, and deliver value to consumers and other interested
parties. [1]
Digital marketing It is also defined as marketing activities that use internet-based media [2]. The internet is a
powerful tool for business. [3] express the characteristics of the internet are as follows:
following:
1. Interactivity, the ability of technological devices to facilitate communication between individuals, such as
face-to-face. The transmission is very interactive so that the
2. participants can communicate more accurately, effectively, and satisfactorily.
3. Demassification, messages can be exchanged to a large number of involved participants.
4. Asynchronous communication technology can send and receive notifications at the desired time of each
participant.
Social media contains information designed by people who use publishing technology to be easily accessible
to facilitate communication, influence, and interaction with others and the general public. The practice of marketing
through social media is starting to develop and is used as a product marketing tool and promoting a company's brand.
Social media is a meeting place for people who want to share information and make new friends, and communicate
with other friends online. Social media, which is developing very rapidly in Indonesia, are Instagram, Facebook,
Twitter, and the like [3].
Social media is a group of internet-based applications that create the ideological and technological foundations of Web
2.0 that allow the creation and exchange of user-generated [4]. Social media allows businesses to reach consumers and build
more personal relationships. [5] divide social media into two groups according to the nature of connection and interaction:
1. Profile-based, namely social media based on profiles that focus on individual members. This group's social media
080004-3
encourages connections that occur because individuals are attracted to those social media users (e.g. Facebook,
Twitter, WhatsApp).
2. Content-based, namely social media that focuses on content, discussions, and comments on the content displayed.
The main goal is to connect individuals with content provided by a particular profile because they like it (e.g.
Youtube, Instagram, Pinterest).
Social media has opened doors for businesses to communicate with millions of people about their products and has
created new marketing opportunities. Utilization of Digital Marketing by MSME Actors Social media can assist MSME
actors in marketing their products [6]. Social media applications are available from instant messaging to social networking
sites that offer users to interact, connect, and communicate. These applications intend to initiate and circulate online
information about the user's experience in consuming a product or brand, with the primary objective of reaching (engage) the
community. In the business context, people engagement can lead to profit creation. [2] found that the digital marketing
strategy had an effect of up to 78% on the competitive advantage of MSMEs in marketing their products. The system consists
of:
1. Availability of product information and product guides; Availability of images such as photos or product
illustrations; Availability of videos capable of visualizing products or presenting supporting presentations;
Availability of attachments for documents containing information in various formats;
2. Availability of online communication with entrepreneurs; Availability of transaction tools and variety of
payment media; Availability of customer assistance and services; Availability of online opinion support;
Availability of testimonial display; Availability of visitor records;
3. Availability of special offers; Availability of the latest information presentation via SMS-blog; Easy product
search; Ability to create brand visibility and awareness; Ability to identify and attract new customers; The
ability to strengthen the brand image received by consumers.
Advantages of Digital Marketing
Digital marketing is a way of marketing that is done by utilizing digital platforms, ranging from social media to
websites. The advantages of using digital marketing include:
a. Ease of Marketing
The first advantage of using digital marketing is the ease of getting an audience with an extensive reach because it is
done on many platforms ranging from popular social media such as Facebook, Instagram and so on. The use of digital
marketing will certainly be able to get more customers with more specific targets.
b. Effective Marketing
Digital marketing also indirectly makes marketing patterns very effective. This is supported by the existence of an
advertising platform that allows companies and business people to target the audience according to the products and services
offered, thus making the costs incurred for marketing needs not in vain.
c. Advertising Costs Are Quite Affordable
It's not enough to stop there; budgeting is another advantage you can get when implementing digital marketing. This is
different from placing advertisements on television or radio media, which require a lot of money and a non-specific target
audience. Digital marketing gives you the freedom to set targets according to your wishes and goals.
d. Data collection
One more advantage that you can get when implementing digital marketing for marketing is the amount of data obtained
to be managed to develop marketing strategies in the future. The data you can get when using digital marketing includes the
keywords you are looking for, gender, age, interests, and the audience's location Disadvantages of Digital Marketing When
discussing the advantages of a strategy, of course, there will be disadvantages, and digital marketing is no exception. Some
of the disadvantages of using digital marketing are:
a. Hard to Run Without Experience
The first drawback in implementing digital marketing is that it cannot be done instantly and requires sufficient experience.
So, as business people and companies, you still need to learn how digital marketing is done step by step. However, almost
all marketing platforms cannot be run instantly.
b. Passion needed
The application of digital marketing must also be balanced with a passion so that its application does not make companies
and business people become unmotivated and tend to be pessimistic. So, identify it first before running digital marketing.
For example, a repair shop mechanic will not feel comfortable working as a mechanic if he is not interested in the automotive
world.
Dissemination of digital marketing strategies in the form of utilizing social media is essential because it can provide
knowledge to MSME actors about the ways and stages in expanding consumer networks through social media in marketing
their products to increase competitive advantage for MSMEs themselves.
080004-4
RESEARCH METHOD
Research Methodology, This study uses a qualitative method approach, by applying a triangulation model, namely
collaborating the method of observation, structured interviews, and in-depth interviews (indepht interviews) and to Micro
Business actors.
Data collection techniques through direct observation and interviews with informants as primary data, while secondary
data comes from literature studies, references, journals, articles, magazines, and the internet.
The sampling technique used by the researcher is purposive sample. Purposive sampling is a sampling technique taking
into account certain conditions [3]. The subject of this research is the owner of goods and services business in regency
totaling 16 SMEs
RESULT AND DISCUSSION
Social media marketing

Social media marketing is a process that encourages individuals to promote their websites, products or services
through online social channels and to communicate by leveraging a much larger community who are more likely to do
marketing than through traditional advertising channels [7]. Social media marketing is a form of online advertising
that uses the cultural context of social communities, including social networks, virtual worlds, social news sites, and
social opinion sharing sites, to meet communication goals [8].
The elements in the application of Social Media Marketing consist of content creation, content sharing, connecting,
and community building [9]. Content Creation can be seen from the creation of exciting content and can represent the
personality of a business so that target consumers can trust it.
1) Content Creation
Engaging content is the cornerstone of a strategy in social media marketing. The content created must be attractive
and must represent the personality of a business to be trusted by target consumers.
2) Content Sharing
Sharing content with social communities can help expand a business' network and grow its online audience. Sharing
content can lead to indirect and direct sales depending on the type of content being shared.
3) Connect
Social networks allow one to meet more people who share the same interests. Extensive networks can build
relationships that can lead to more business. Honest and careful communication must be considered when doing social
networking.
4) Community Building
Web Social media is a large online community of individuals where interactions occur between people living
worldwide using technology. Building a community on the internet that has the same interests can happen with social
networking.
Identified the business value of using social media for MSMEs, such as the creation of sustainable marketing
channels, increased revenue, decreased advertising costs, reduction in marketing costs, a result of competitive
advantage, ease of promotion across social media platforms, increasing the popularity of the brand, and the
introduction of the company to the public [4]. The Micro, Small and Medium Enterprises sector has characteristics
that can distinguish it from large-scale enterprises. In the perspective of its development, according to [10], SMEs can
be grouped into 4 (four) classifications, as follows:
1. Livelihood Activities are MSMEs that are more commonly known as the informal sector.
2. Micro Enterprise is an MSME that has the nature of craftsmen but still does not have entrepreneurial
characteristics.
3. Small Dynamic Enterprise is an MSME that has an entrepreneurial spirit and can accept subcontract and
export work.
4. Fast Moving Enterprise is an MSME that already has an entrepreneurial spirit and will seek transformation
into a Big Business (UB).
The use of the internet has changed most marketing concepts, which previously only relied on the marketing mix
consisting of product, price, place, and promotion. Still, the current development is an addition to the basis of
marketing, namely the 4C approach consisting of the customer, cost, convenience and communication [11]. Businesses
must consider charges to maximize profits and use the internet to make consumers comfortable because consumers
can buy products only from home. Companies need to develop two-way communication to form a harmonious
080004-5
relationship (goodwill) with customers [11].
Data analysis through the process of data reduction, data display, and verification or conclusion.
Based on the information presented to the SMEs participants, the Regency SMEs business actors implement the
use of social media applications to expand the marketing network for the long term by paying attention to the
elements in the implementation of the implementation in stages
Characteristics of MSMEs
The subjects in this study are Regency MSME actors who produce or sell various types of goods and food with a
total of 16 MSMEs Beginners; these MSMEs during the pandemic tried to rise by trying to make breakthroughs
through digital marketing using several social media applications. The details of MSMEs who become informants are
in table 1
Use of Social Media in Regency MSMEs

Based on the results of data collection conducted through the effects of in-depth open questionnaires and
observations with informants who are MSMEs in the regency, it is known that digital marketing by utilizing social
media, social chat, marketplace and other platforms makes it easier for business actors to attract and interact directly
with consumers. Technological developments make business actors need to adapt, especially in the marketing process.
Of the 16 MSME businesses who became informants in this study, they stated that the use of information technology
by utilizing the internet made them feel helped, moreover, the tendency of consumers today to be very active on social
media and prefer to shop online rather than going to shops, markets, supermarkets, or malls.
The role of social media that is getting better in supporting MSME activities is not missed by SMEs after gaining
knowledge about digital marketing. As for social media, that is often used by MSME actors are WhatsApp (WA) by
50% and Instagram (IG) by 18.75% for product images. SMEs actors update the information displayed as business
information and communication to customers. (table 2).
The Effect of Social Media Use in Regency MSMEs

The results obtained from an open questionnaire to informants who are Regency MSMEs actors based on the 4C
approach (cost, customer, convenience, and communication) agree that digital marketing helps them promote and
market their products effectively and efficiently. Informants argue that digital marketing saves on promotional costs
because they still rely on offline marketing such as distributing brochures, advertising in newspapers, and advertising
through billboards; their expenses are more significant when using offline media. Previous research also reinforces
this, which states that internet marketing expands market reach, lowers marketing costs, and improves good
relationships with customers [12].
Several informants agreed that digital marketing provides cost savings, increases interaction with customers and
provides convenience. Some of the things that informants do are updating information through social media, websites,
marketplaces, special customer groups, etc. On average, they update info once a day to retain old customers and attract
new consumers to buy their products. With the internet, they can save costs for promotion. When they use social
media, they only need an internet quota and those with a particular budget for advertising through digital. They can
take advantage of Facebook, Instagram and Whatsapp because they are the most recognized platforms and others.
Some consumers state that promoting through digital increases their sales; this is influenced by the ease with which
consumers access company information and order products.
The results of using social media applications for a short time have not produced anything very significant. Based
on the data in the field after the joint survey, it was found that 42% of them had a fixed turnover, 0.8% increased, and
50% experienced a decline. (table 3).
For average turnover, use one social media application, both WhatsApp and Facebook. It is necessary to understand
the tricks when it is the right time to do promotions to increase turnover. MSMEs use two or three social media
applications to increase turnover, both Whattshap, IG and Facebook. The strategy in advertising has been good enough
for SMEs. For turnover to fall, SMEs use one social media application. Namely, what shape the business. The
promotion strategy carried out by SMEs has not been maximized; it is necessary to combine applications to get a much
better turnover
Based on the research results on the use of applications, it is generally in line with the study conducted by Syed
Zamberi Ahmad, Abdul Rahim Abu Bakar, and Norita Ahmad (2019) [13] in the journal-title Social media adoption and
its impact on firm performance: the case of the UAE. However, SMEs have not tried all applications simultaneously
because they do not know how to operate the technology. Although social media has not been fully utilized, SMEs
080004-6
actors, especially in the regency, have used other digital media to process and promote their products.
FIGURES AND TABLES
Figure 1 shows that Indonesia's total population or population reaches 272.1 million with access via mobile of 338.2
million, through the internet of 175.4 million, and through active social media of 160 million.
Figure 2. shows that internet usage via cellphones reached 171 million with access via cellphones by 98% of the
total users, 96% using cellphones for data transfer, 5.3% using features or applications on cellphones for data transfer
and average cellphone usage per day of 4 hours 45 minutes.
Figure 3 shows that online product search activities make purchases by 93%, visit online stores by 90%, order
products online as much as 88%, make online orders through laptops by 25%, and make online orders via HP by 80%.
Figure 4. shows that the most widely used social media applications are YouTube 88%, WhatsApp 84%,
Facebook 82%, Instagram 79% and Twitter 56%.
FIGURE 1. internet and social media trends 2020 in Indonesia (Hootsuite) [14]
080004-7
FIGURE 2. Internet Usage (Hootsuite) [14]
FIGURE 3. Electronic commerce activities (Hootsuite) [14]
080004-8
FIGURE 4. Social media platforms (Hootsuite) [14]
Table 1. Regency SMEs
No BUSINESS NAME TYPE OF BUSINESS
1 HANDAYANI SELLING WET FOOD

KITCHEN
2 POP CHICKEN JUL SNACK
3 NATHALIE KITCHEN HOME CATERING SERVICES
4 LAUNDRY NANA LAUNDRY SERVICE
5 FLORIST SELLING FLOWER
6 TV PARTNER CABLE TV SERVICES
7 GHEA SHOP CLOTHING IRONING SERVICES
8 AIDIL SHOP Mixed Flour and Salted Eggs
9 FAITH SHOP SELLING WET FOOD
DAOEN YELLOW
10 SELLING FOOD
RICE
11 SILVIA PEMPEK SELLING FOOD
12 ALI SELLING SHOES
13 SHOP DIAZ SELL FOOD
14 FRIED BANANAS SELLING FRIED BANANA VARIOUS FLAVOR
15 SALON HOME SALON
16 AM CARGO DELIVERY SERVICE
080004-9
Table 2. Social media used by Regency SMEs
NO APPLICATION TOTAL PERSENTATION

1 FACEBOOK 9 56,25%
2 WHATSAPP 8 50,00%
3 INSTAGRAM 3 18,75%
4 TWEETERS 0 0,00%
5 OTHERS 4 25,00%
Table 3. Regency SMEs Turnover
BUSINESS TURNOVER WITH

No BUSINESS NAME DIGITAL MARKETING
PERMANENT RIDE DOWN
1 HANDAYANI V
KITCHEN
2 POP CHICKEN V
3 NATHALIE KITCHEN V
4 LAUNDRY NANA V
5 FLORIST V
6 TV PARTNER V
7 GHEA SHOP V
8 AIDIL SHOP V
9 FAITH SHOP V
DAOEN YELLOW
10 V
RICE
11 SILVIA PEMPEK V
12 ALI V
13 SHOP DIAZ V
14 FRIED BANANAS V
15 SALON V
16 AM CARGO V
TURNOVER TOTAL PERSENTATION

PERMANENT 6 37,50%
RIDE 2 12,50%
DOWN 8 50,00%
080004-10
CONCLUSION
Users of digital technology have influenced all aspects of human life and activities, including marketing.
Digital-based marketing is used to acquire consumers, build their preferences, promote brands, nurture consumers,
increase the number of products sold and ultimately increase profits. Digital Marketing allows buyers to get all
information about products and make buying and selling transactions online. Like sellers, they can also monitor and
meet the needs and wants of potential buyers regardless of time and geographic attachment to specific products and
brands.
Changes in marketing behaviour from conventional to digital are not matched by the presence of MSME actors
who use digital marketing, even though MSMEs are believed to be able to spur the Indonesian economy. The use of
the digital marketing concept is hope for MSMEs to develop into economic powerhouses. MSMEs can take advantage
of digital technology, especially social media, as a means of digital marketing. Besides being low-cost, and not requiring
special skills in initial initiation, social media is considered capable of reaching consumers directly. The lack of
knowledge of MSME actors about digital marketing has inspired this community service team to become an obstacle
for business people in marketing their products over time.
Based on the research results on the use of applications in pandemic that is ride 12,5%, Permanent 37,5% and
doen 50%, it is generally in line with the study conducted by Syed Zamberi Ahmad, Abdul Rahim Abu Bakar, and
Norita Ahmad in the journal-title Social media adoption and its impact on firm performance: the case of the UAE. The
utilization of technology at Regency MSMEs has not been maximized. However, the enthusiasm of MSME actors is
interested in using digital marketing and utilizing social media optimally but faces several obstacles, such as a lack of
understanding of information technology. Their knowledge of maximizing the Youtube platform WhatsApp, Facebook,
Instagram, Marketplace is very lacking, making exciting and interesting "posts".
ACKNOWLEDGMENTS
We would like to thank all the support given in the completion of this research. Hopefully, this research can be
helpful as an effort to increase the knowledge of the marketing family.
REFERENCES
[1] P. K. Kannan and H. “Alice” Li, “Digital marketing: A framework, review and research agenda,” Int. J. Res.
Mark., vol. 34, no. 1, pp. 22–45, Mar. 2017, doi: 10.1016/J.IJRESMAR.2016.11.006.
[2] A. Wardhana, “Strategi Digital Marketing Dan Implikasinya Pada Keunggulan,” Prosidinr Semin. Nas. Forum
Keuang. dan Bisnis, no. April 2015, pp. 327–337, 2018, [Online]. Available:
https://www.researchgate.net/publication/327069950%0ASTRATEGI.
[3] A. Martoyo, “Adopsi Media Sosial Sebagai Sarana Pemasaran Digital Bagi Usaha Mikro Makanan dan Minuman
Di Kota Jakarta Selatan,” … , J. Ilm. Ilmu Manajemen, Bisnis dan …, vol. 1, pp. 1–14, 2020, [Online]. Available:
https://ejournal.bri-institute.ac.id/index.php/bricommerce/article/view/49.
[4] I. Paul and S. Ganguly, “Bovine Mastitis, an Economically Important Bacterial Infection of Udder in Cattle: a
Review,” Ind. J. Sci. Res. Tech, vol. 2, no. 2, 2014, [Online]. Available: http://www.indjsrt.com.
[5] Z. Qingyun and J. Sarkis, “Green Marketing and Consumerism in China: Analyzing the Literature,” Cent.
Sustain. Bus., vol. 23, no. 3, pp. 230–239, 2015.
[6] M. Yadav, “Social media as a marketing tool: Opportunities and challenges,” Indian J. Mark., vol. 47, no. 3, pp.
16–28, 2017, doi: 10.17010/ijom/2017/v47/i3/111420.
[7] M. A. Rahman, “Studi Penerapan Digital Marketing Pada UKM Makanan Khas di Kota Malang,” Artik. Ilm. Ed.
40 Tahun IV-Januari 2020, pp. 46–50, 2020.
[8] B. Zenelaj and V. Gambarov, “Marketing of Political Parties on Social Media,” 5th Int. Conf. Eur. Stud., pp.
225–242, 2015.
[9] A. Studi, K. Pt, M. Padi, and A. Manajemen, “Analisis sosial media marketing dalam membangun,” vol. 7, no.
2, pp. 4303–4312, 2020.
[10] D. Krisnawati, “PERAN PERKEMBANGAN TEKNOLOGI DIGITAL PADA STRATEGI PEMASARAN
DAN JALUR DISTRIBUSI UMKM DI INDONESIA (Studi Kasus: UMKM Kuliner Tanpa Restaurant
080004-11
’Kepiting Nyinyir’),” vol. 6, no. 1, pp. 2338–4794, 2018.
[11] K. T. Smith, “the Marketing Mix of Imc: a Move From the 4 P’S To the 4C’S,” J. Integr. Mark. Commun., pp.
2–7, 2003.
[12] M. R. Abbas, A. O. Abdullateef, and S. S. M. Mokhtar, “Does Service Recovery Lead to Customer Satisfaction
and Loyalty in Airline Industry ? A Perceived Justice Theory Approach,” World Appl. Sci. J., vol. 33, no. 2, pp.
256–262, 2015, doi: 10.5829/idosi.wasj.2015.33.02.960.
[13] S. Z. Ahmad, A. R. Abu Bakar, and N. Ahmad, “Social media adoption and its impact on firm performance: the
case of the UAE,” Int. J. Entrep. Behav. Res., vol. 25, no. 1, pp. 84–111, 2019, doi: 10.1108/IJEBR-08-2017-
0299.
[14] K. Simon, “Digital 2020: Indonesia,” 2020, [Online]. Available: https://datareportal.com/reports/digital-2020-
indonesia.
080004-12
An International Tourist Behaviour on Mobile Smartphone
Usage
Jack Febrian Rusdi1,2, a), Nur Azman Abu2, b), Sazilah Salam2, c), Harya Gusdevi1, d),
Richki Hardi3, e), and Daniel Guna Nugraha4, f)
Author Affiliations
1
Teknik Informatika, Sekolah Tinggi Teknologi Bandung, Bandung, Indonesia.
2
Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Melaka, Malaysia.
3
Teknik Informatika, Universitas Mulia, Balikpapan, Indonesia.
4
Associations of Indonesian Tours and Travel Agencies (ASITA), Head Office, Jakarta, Indonesia.
Author Emails
a)
b)
nura@utem.edu.my
c)
sazilah@utem.edu.my
d)
devi@sttbandung.ac.id
e)
f)
tours@exoticjavatrails.com
Abstract. Tourism plays an essential role in the world economy. Drastic progress in this field has reshaped an integral
part of tourism information and technological communication supports. Mobile devices have consumed significant
aspects of people's daily life. Understanding tourist behaviour is considered an essential element to the success of
development in tourism. An international tourist no longer relies solely on a tourism agency to determine their travel
plans. This paper concentrates on specific behaviours of international tourists related to their interaction with mobile
devices. Methods used in this research are questionnaires and literature study. The questionnaire was conducted in 2019
before the pandemic. This research result is which media are being used in determining travel plans, mainly how long
mobile devices are being used, the social media chosen, and the types of activities involved in using these devices. This
research can be a significant base on embarking on a potent strategy in the tourism industry.
1. INTRODUCTION
Tourist travel activities grow continuously. Tourism contributes 10.4% of the Gross Domestic Product in the
world. There were 8.8 trillion USD generated through the tourism sector in 2018 alone. During 2018, there have
been 1.326 million international tourists from various countries. Tourism has contributed a significant impact on the
world economy. This sector creates jobs, drives export value and prosperity between countries (UNWTO, 2018;
WTTC, 2019).
To better feed into information needs, an investigation on the behaviour of international tourists has been done
and focusing on the following:
1. Identify factors that influence tourist decisions in choosing a tourist destination
2. Know the time used related to the use of smartphones for tourists when travelling
3. Know how tourists are connected to the internet when travelling
4. Understanding the social media used by tourists when travelling
5. Identify tourist activities related to smartphone use when travelling
080005-1
As technology research related to tourist behaviour (Rusdi, Salam, Abu, Baktina, et al., 2021), this research is the
basis for the development of various tourism strategies, including implementation of GIS with mapping (Lahayon et
al., 2020; Shanono et al., 2019; Sunaryo et al., 2019), behaviour tracking (Hardi et al., 2021; Rusdi, Salam, Abu,
Sahib, et al., 2019; Rusdi, Salam, Abu, Sunaryo, Naseer, Rismayadi, et al., 2021; Rusdi, Salam, Abu, Sunaryo,
Naseer, Setiawan, et al., 2021; Taufiq et al., 2020), reporting system (Hendayun et al., 2021; Rusdi, Salam, Abu,
Sunaryo, Agustina, et al., 2021), usage of vision technology, and the development of tourism instructors in various
aspects, including aviation (Setiawan et al., 2020), system security (Naseer et al., 2021). This research can also be
used as a basis for improving skills, especially in tourism vocational education (Ariyani et al., 2021).
METHODOLOGY
Data collection has been arranged by the questionnaire method. The method stage of this research consists of 3
stages, namely preparation, execution, and review (Rusdi, Salam, Abu, Sunaryo, et al., 2019).
• Identification of Information Needs

• Classification of tourist types and number of target participants
• Make an questionnare Guide Form
Preparation
n • Data Obtaining Strategy
• Approach to Tourists
• The Process of Filling in the Form
• Collection of results
Executionn • Selection of collected data and filing process
• Converting manual data to electronic data

• Data Grouping
Review • Make a report on results
FIGURE 1. Research Stage
At this stage, tourist information has been identified on tourist behaviours related to the use of information
technology which includes tourist demographics, identification of sources of information that affect tourists in
planning trips, how tourists get internet access while on vacation, social media that tourists use, and smartphone
functions for tourist when travelling (see FIGURE 1).
FIGURE 2. Topics of Questionnaire
The classification of topics is extracted from international tourists who do not involve local or national tourists as
participants of this study. Please refer to FIGURE 2 for the subject of this questionnaire. The minimum targeted
number of tourists is from 40 countries and 250 participants.
The questionnaire guidelines have been created based on the information that needs extracting. The guidance is
made in the format of one soft paper ready to be filled by tourists, both in essays and optional questions, and closed
by the tourist signature section. A tourist questionnaire survey has been conducted in Bandung, Indonesia
(Desfitriady, 2020; Rusdi, 2017; Tanjung et al., 2021).
080005-2
The strategy to get data is to meet tourists in the field, especially in tourist attractions that are easy to find, such
as international restaurants; natural tourist attractions such as crater and hot spring; heritage tourism sites; shopping
places frequented by tourists; Hotel; educational places such as universities with educational programs for
international students, and international schools. Also, by executing directly onto the field, researchers collaborate
with tour service companies.
At the execution stage, an approach to tourists is carried out by introducing the aims and objectives of the
research and asking tourists to fill in the forms provided (Rusdi, Salam, Abu, Sunaryo, et al., 2019). The
questionnaire with tourists was carried out from 11 April - 28 May 2019. Out of 306 tourists encountered, 302 data
were used as the basis of this study (n = 302). One person refused for privacy reasons, while three other tourists did
not complete the demographic data, including questions left blank (TABLE 1).
TABLE 1. Respondents' demographic profile

Gender Education
Country Origin Total
M F BU D/UG PG Other
Australia 25 21 4 20 5 17 46
Malaysia 20 26 0 31 9 6 46
Netherlands 22 14 4 11 11 10 36
US 16 10 3 10 5 8 26
Singapore 10 14 4 7 3 10 24
France 5 7 1 4 2 5 12
UK 5 4 0 6 0 3 9
Germany 1 7 3 2 1 2 8
Philippines 3 5 0 3 5 0 8
New Zealand 7 0 0 0 0 7 7
Canada 4 2 1 3 1 1 6
China 2 4 0 3 3 0 6
Italy 1 4 3 1 1 0 5
Pakistan 3 2 1 0 3 1 5
Poland 2 2 0 0 1 3 4
England 2 1 0 2 0 1 3
Greece 3 0 0 3 0 0 3
Japan 1 2 0 3 0 0 3
Belgium 1 1 0 0 2 0 2
Brunei Darussalam 1 1 0 2 0 0 2
Ireland 1 1 0 0 2 0 2
Kazakhstan 0 2 0 0 1 1 2
Latvia 0 2 0 1 1 0 2
Mauritius 2 0 0 2 0 0 2
Saudi Arabian 2 0 0 2 0 0 2
Spain 1 1 0 1 1 0 2
Swiss 1 1 0 0 2 0 2
Switzerland 1 1 0 0 1 1 2
Taiwan 0 2 0 1 0 1 2
Tanzania 2 0 0 0 2 0 2
OTHER 13 9 0 7 6 9 22
TOTAL 156 146 24 124 68 86 302
Note: BU = Below University; D = Diploma; UG = Under Graduate; PG = Post Graduate
080005-3
Furthermore, collected data is allocated to the database at the review stage, grouped according to needs, and
research reports are made based on the database process. Statistical analysis in this study was carried out with a
percentage analysis according to the type of data produced. The limitation of the results of this study is carried out
and displayed only by percentage analysis, and this research did not use for high-level statistical purposes.
FINDINGS
Participants came from 52 countries. Demographic profiles are shown in TABLE 1—Source of information for
travel planning. The collected data revealed that tourist decision-making is influenced by various details from the
media or recommendations during the trip planning stage. Suggestions come from friends, family, colleagues,
companies, and part of group trips. This recommendation can be related to the experience of other tourists or
purposes because of business, work, or education.
Resource of Information (%)
Social Media 64,90%
Recommendation 32,78%
Online Advertising 32,45%
Blog 22,85%
Tripadvisor 21,85%
Search Engine 19,54%
TV 18,87%
Magazine 14,24%
Advice Agent 8,28%
Newspaper 8,28%
FIGURE 3. Resource of Information
Out of ten options for information sources in travel planning for tourists, 24 tourists (7.95%) did not fill out
information related to this recommendation. Tourists who filled one choice were 85 people (28.15%), two choices
were 53 people (17.55%), three choices were 65 people (21.52%), four choices were 37 people (12.25%), five
choices were 29 people (9.60%), six choices of 5 people (1.66%), seven choices of 4 people (1.32%) and no one
gave more than seven choices (see FIGURE 3).
Based on input submitted by tourists, references can come from family, friends, and assignments both for study
and business.
Internet Connection by Tourists (%) Tourist smartphone usage while traveling (hours)
Philippines 10,38
Tourist Destination 6,62%
Malaysia 7,74
New Zealand 7,43
Restaurant 43,71%
Germany 7,00
US 6,57
Hotel 77,81%
Australia 5,79
Roaming 7,62% UK 4,88
Singapore 4,73
Local Operator 46,36% France 4,00
Netherlands 3,34
FIGURE 4. (a). The way tourists access the internet via a smartphone, (b) smartphone usage in hours.
080005-4
Internet access when travelling. When travelling, there are five ways revealed by tourists to be able to connect to
the internet, namely via Wi-Fi available at hotels, restaurants, and tourist destinations — besides, connection to the
internet made through their cellular telephone network, both roaming and local operators. A total of 163 people
(53.97%) use internet access through their mobile cards connected to the mobile telephone network. Please see this
finding in FIGURE 4a.
Tourist smartphone usage while travelling in hours. Overall, the average number of tourists using a smartphone
in one day is 6.45 hours. Of the ten highest tourist origin countries, the accumulated duration of smartphone usage
during travelling saw in FIGURE 4b.
The use of social media for tourists when travelling. Regarding the use of social media by tourists when
travelling, as many as 48 tourists (15.9%) did not mention the social media they used, and 3 of them said that they
did not use any social media on their trip abroad.
Based on the data collected, as shown in FIGURE 5a, the information displayed is for the use of social media
for tourists with a value above 1%. Whereas social media with low values are not displayed on the graph, they are
Booking, Camera, Pinterest, Skype, BCA, BitCoin, Gojek, Grab, GVS, Hello, IMO, LOL. Mail, Maps, Pal, PB,
PUBG, Spotify, Tik Tok, Tinder, Video Call, WeChat. Weibo and Wi-Fi.
Usage of Social Media by Tourists (%) Smartphone Function by Tourist
IG 63,91 Taking photos 87,09%
FB 54,3
Maps feature 71,52%
Telephone 62,25%
WA 40,73
Media posting 52,32%
Twitter 11,92 Restaurant search 51,66%
Youtube 4,64 Share photos 50,66%

Video Call 45,70%
Line 3,31
Translator 41,06%
Gmail 2,65
Search of attraction 41,06%
Snapchat 2,65 Reading news 34,77%
Telegram 2,32 Online Banking 34,44%

Whatsapp 28,15%
Google 1,99
Currency Converter 28,15%
LinkedIn 1,66
As tour guide 11,92%
TA 1,32
Path 1,32
FIGURE 5. (a) Usage of social media by tourists in percentages, (b) Smartphone function by the tourist while travelling
Smartphones function by the tourist while travelling. On what activities are carried out by this international
tourist with their smartphone devices while travelling, a total of 6 tourists (1.99%) did not explain or choose any
activities they did. Tourist activities during travel related to smartphone use, as shown in FIGURE 5b
DISCUSSION
In preparing travel plans for tourists (FIGURE 3), at the pre-trip stage, it was seen that most tourists (64.90%)
received information from social media, then referenced and online advertising. In preparing travel plans, we
noticed that the influence of the existence of references from Information Technology is a potential means for
tourists to arrange their travel plans, such as social media, Online Advertising, Blogs, TripAdvisor, and Search
Engines. Other influences that we need to consider are recommendations, given that this source was delivered by as
many as 32.78% of the participants studied.
More than half of tourists (53.97%) connected to the internet via mobile operators, roaming, and using local
operators (Error! Reference source not found.a). Data revealed that as many as 77.81% of tourists access the internet
from hotels. This research result shows that most international tourists still need to be connected to the global world
when travelling and are higher when they are already at the hotel. In general, the point of view of smartphone device
use (FIGURE 4b), the dependence of tourists on these devices when travelling is relatively high. On average of all
participants daily use mobile phones for 6.45 hours.
080005-5
Regarding the use of social media for tourists (FIGURE 5a), the three social media rankings used are
Instagram, Facebook, and WhatsApp. Instagram is the highest, reaching 63.91% of the total tourists studied. This
research also reveals the behaviour of tourists related to the use of their smartphone devices, as we saw from the
function of the device during travelling (FIGURE 5b). The smartphone is more likely to be used as a camera to
document travel. Its use as a camera is relatively high, with as many as 87.09% of all participating tourists.
The function of smartphones used by more than half of the international tourists who provide information
through this research is related to maps features, telephones, posts on social media, and various photos.
CONCLUSION
Information Access has become an integral part of daily life. Information consumption by a tourist via a
smartphone has expanded considerably in the presence of a mobile network. Usually, tourists will utilize and depend
highly on their camera facilities, logistical map, and social media available on their mobile devices. Consequently,
international tourists have embarked on new outlooks and behavioural dependence on mobile technology.
Considering these new behaviours and addiction on a person's smartphone at their fingertips, the relevant
stakeholders in the tourism industry should develop and build new approaches and business strategies through
advanced technology to gain a competitive advantage in a global tourist market.
Further research needs to explore the following: how is the behaviour of tourist movements through location-
based technology? Can the use of telephone calls through data packets generally replace telephone calls through
conventional methods for tourists? Is a connection to the global world a vital thing when tourists are at tourist
destinations? Moreover, various forms of technological development can facilitate tourists associated with their
behaviour in using technology, both the development of algorithms, methods, structures, strategies, devices,
software and others.
ACKNOWLEDGMENTS
This research was conducted by the Pervasive Computing & Educational Technology Research Group. C-ACT,
Universiti Teknikal Malaysia Melaka (UTeM). The Indonesia Tourism Journalist Association (ITJA) and
Associations of Indonesian Tours and Travel Agencies (ASITA) have provided access to various parties related to
this research. Sekolah Tinggi Teknologi Bandung and PT Jackwisata (jacktour.com) have provided research
materials related to technology and tourism.
REFERENCES
Ariyani, L. F., Widjaja, S. U. M., Wahyono, H., Haryono, A., Rusdi, J. F., & Pratama, C. B. A. (2021). Vocational education
phenomena research method. MethodsX, 8, 101537. https://doi.org/10.1016/J.MEX.2021.101537
Desfitriady. (2020). Indonesia: Regional Deconcentration Management. SciTech Framework, 2(1), 34–42.
Hardi, R., Suryana, N., Pee, N. C., Pribadi, A. S., Rusdi, J. F., & Junaidi, A. (2021). The Object Tracking System at the Service
Delivery Center of the Traveling Salesperson Problem Method. Journal of Physics: Conference Series, 1807(1), 12034.
https://doi.org/10.1088/1742-6596/1807/1/012034
Hendayun, M., Yulianto, E., Rusdi, J. F., Setiawan, A., & Ilman, B. (2021). Extract transform load process in banking reporting
system. MethodsX, 8. https://doi.org/10.1016/j.mex.2021.101260
Lahayon, G. H., Esma, G. T., Sinday, L. C., Juventud, J., Pitogo, V. A., & Febrian Rusdi, J. F. (2020). GENEALOGICAL
AGGREGATION AND ANCESTRAL DOMAIN MANAGEMENT: A WEB-BASED GIS IMPLEMENTATION IN A
DEVELOPING COUNTRY. 41st Asian Conference on Remote Sensing, ACRS 2020, 9. https://a-a-r-
s.org/proceeding/ACRS2020/1qbggq.pdf
Naseer, M., Rusdi, J. F., Shanono, N. M., Salam, S., Muslim, Z. bin, Abu, N. A., & Abadi, I. (2021). Malware Detection: Issues
and Challenges. Journal of Physics: Conference Series, 1807(1), 12011. https://doi.org/10.1088/1742-6596/1807/1/012011
Rusdi, J. F. (2017). Kolaborasi Pariwisata Bandung Raya. Indonesia Tourism Journalist Association.
Rusdi, J. F., Salam, S., Abu, N. A., Baktina, T. G., Gumilar Hadiningrat, R., Sunaryo, B., Rusmartiana, A., Nashihuddin, W.,
Fannya, P., Laurenty, F., Shanono, N. M., Hardi, R., Kuswayati, S., Damayanti, S. E., & Rahmawati, S. (2021).
Collaborative of ICT Research in Indonesia. Journal of Physics: Conference Series, 1807(1), 12009.
https://doi.org/10.1088/1742-6596/1807/1/012009
Rusdi, J. F., Salam, S., Abu, N. A., Sahib, S., Naseer, M., & Abdullah, A. A. (2019). Drone Tracking Modelling Ontology for
Tourist Behavior. Journal of Physics: Conference Series, 1201(1), 012032. https://doi.org/10.1088/1742-
6596/1201/1/012032
080005-6
Rusdi, J. F., Salam, S., Abu, N. A., Sunaryo, B., Agustina, N., Gusdevi, H., Hermansah, Hadhiwibowo, A., Heryati, T., Hardi, R.,
Taufiq, R., Jufri, M. T., Sudarsono, I., Utomo, E. W., Kodong, F. R., & Fannya, P. (2021). Reporting of Hospital Facility
on Smartphone. Journal of Physics: Conference Series, 1807(1), 012013. https://doi.org/10.1088/1742-
6596/1807/1/012013
Rusdi, J. F., Salam, S., Abu, N. A., Sunaryo, B., Naseer, M., Rismayadi, D. A., Kodong, F. R., Sudarsono, I., Utomo, E. W.,
Yuhefizar, Pitogo, V. A., & Shanono, N. M. (2021). Field Reporting Irrigation System via Smartphone. Journal of
Physics: Conference Series, 1807(1), 12012. https://doi.org/10.1088/1742-6596/1807/1/012012
Rusdi, J. F., Salam, S., Abu, N. A., Sunaryo, B., Naseer, M., Setiawan, A., Kchouri, M., Taufiq, R., Respati, R. D., Alam, S. N.,
Sarah, S., Rusmartiana, A., & Rahmawati, S. (2021). A Tourist Tracking Model by Tourist Bureau. Journal of Physics:
Conference Series, 1807(1), 12010. https://doi.org/10.1088/1742-6596/1807/1/012010
Rusdi, J. F., Salam, S., Abu, N. A., Sunaryo, B., Taufiq, R., Muchlis, L. S., Septiana, T., Hamdi, K., Arianto, Ilman, B.,
Desfitriady, Kodong, F. R., & Vitianingsih, A. V. (2019). Dataset Smartphone Usage of International Tourist Behavior.
Data in Brief, 27(December 2019), 104610. https://doi.org/10.1016/j.dib.2019.104610
Setiawan, A., Yulianto, E., Abadi, I., & Ginanjar, A. (2020). Obstacles Analysis Digitalization of Flight Area With Aerodrome
Elevation System Approach. SciTech Framework, 2(1), 1–11.
Shanono, N. M., Kodong, F. R., & AL-Jaberi, M. A. A. (2019). Monitoring Infectious Diseases Diffusion through GIS. SciTech
Framework, I.
Sunaryo, B., Rusydi, M. I., Rusdi, J. F., Suriani, R., & Daus, S. (2019). Sistem Pelacakan Lokasi Pelaporan Petugas Lapangan
Irigasi Provinsi Sumatera Barat Berbasis GPS Smartphone dan WebGIS. Jurnal RESTI, 3(2), 271–281.
https://doi.org/https://doi.org/10.29207/resti.v3i2.957
Tanjung, A. H., Salam, S., Rusdi, J. F., Ermawati, Y., Novianty, I., Hendaris, R. B., & Apriliawati, Y. (2021). Flypaper effect
assessment methods in the expansion of regional autonomy. MethodsX, 8, 101387.
https://doi.org/10.1016/j.mex.2021.101387
Taufiq, R., Baharun, M., Sunaryo, B., Pudjoatmodjo, B., & Utomo, W. M. (2020). Indonesia: Covid-19 and E-Learning in
Student Attendance Method. SciTech Framework, 1(1), 12–22.
UNWTO. (2018). Tourism Highlights 2018 Edition. In UNWTO Tourism Highlights: 2018 Edition. UNWTO.
https://doi.org/10.18111/9789284419685
WTTC. (2019). Travel & Tourism: Economic Impact 2019.
080005-7
Comparison of Support Vector Machine and K-Nearest
Neighbor Algorithms in Indonesia Hoax Classification
Inte Christinawati Bu’ulolo1,a), Ivan Felix Sinaga1,b), Wira Epriana Ambarita1,c)
Author Affiliations
1
Del Institute of Technology, Jl. PIDel, Sitoluama, Laguboti, Toba Regency, North Sumatera, Indonesia
Author Emails
a)
Corresponding author: inte@del.ac.id
b)
ivansinaga372@gmail.com
c)
ambaritawira@gmail.com
Abstract. Today, more and more people spread the word and read news from social media. This is because on social
media it is easier to spread the news and to get news. So, to overcome these problems, it is necessary to classify the news
contained on social media whether the news is hoax news or facts because it is important to assess the level of truth of a
news spread on social media. To classify which news is hoax or fact, manually will spend a lot of time, so one solution to
make it easier to classify hoaxes or facts is to use a machine learning approach. In this study, researchers used the SVM
and KNN algorithms to classify hoax news in Indonesian. From the two algorithms, we will look for which model is the
best in classifying hoax news. Researchers use four categories of data that have their own characteristics, in this study
several stages were carried out starting from the preprocessing stage and using TF-IDF calculations in measuring the
weight of each news, then the learning process was carried out using the Support Vector Machine algorithm and K-
Nearest Neighbor and evaluation of the model by using the confusion matrix. Based on the results of the model
evaluation, it was found that the classification with the SVM algorithm obtained a higher accuracy value than KNN, with
an accuracy value of 84% and f-measure of 84% in the third data category.
Keywords: Machine learning, data preprocessing, Support Vector Machine, K-Nearest Neighbor, confusion
matrix, accuracy, f-measure
PRELIMINARY
In this millennial era, often in their daily life is spent interacting online through social media, and more and more
people tend to spread and consume news from social media rather than news media such as newspapers, television,
or radio. The factor that causes people to use social media more often is because they can disseminate information in
minutes to millions of people and get news cheaper and on time. The spread of fake news is growing rapidly in line
with the development of information technology which causes everyone to post news at will and social media has
proven to be a powerful source for the spread of fake news. From the problems discussed above, it is necessary to
research in the detection of hoax news on social media. This research is important because social media users are
faced with a lot of information every day, and it is impossible to judge the credibility of each piece of information to
know whether the news is fake or fact.
In this study, to classify hoax news, researchers used a supervised learning algorithm for classification, namely the
Support Vector Machine algorithm and the K-Nearest Neighbor algorithm. Support vector machine is a machine
learning algorithm that uses linear constraints for the base so that it can process various kinds of data [1]. SVM is a
supervised algorithm for classification that works by finding the hyperplane with the largest margin. However, in the
application of the SVM algorithm, not all data can be separated linearly in two dimensions, so that the linear
constraint is converted into hyperplanes and by utilizing the hyperplanes kernel function can separate data in higher
dimensions [1]. K-Nearest Neighbor (KNN) is a technique of lazy learning and KNN is included in instance-based
learning, the KNN algorithm will look for k object groups in the training data that are closest to the new object in the
training data. The KNN algorithm will classify objects based on the closest (similar) distance to the learning data
with the object.
090001-1
In this study, researchers used a dataset, namely Indonesian news taken from 2017-2019. The dataset is divided
into four categories. In this study, there are four categories of datasets used because the researcher wants to see how
the classification is done using the four categories of data where the first category is a scrapping dataset from the
website turnbackhoax.id/, the second category of datasets is the modification of the first data by reducing the
number of hoax data, the third category is the modification of the first category data by adding factual data from
other sources such as cnidonesia.com , detik.com , tempo.co , coverage6.id, and the fourth category is mirroring
data, which is a dataset where hoax news and facts have a relationship (the news topics are related). The use of the
first category, because the researcher wants to know how the scrapping data taken automatically from the website is
used in the classification of hoaxes and facts. The use of data in the second and third categories, because researchers
want to know how the balance of the spread of hoax data and facts and the amount of data affect the classification.
The use of the fourth category is because researchers want to know how related data, namely news data, have the
same topic between hoax news and influential facts in the classification results.
The classification of hoax news and facts in this study uses TF-IDF and does not use the semantic method and the
results of the classification of hoax news in this study do not include grammatical errors. The research focuses on
analysis to get the algorithm with the highest accuracy in the classification of hoaxes. The purpose of this research is
to obtain a classification model for hoax news or facts using the Support Vector Machine and K-Nearest Neighbor
algorithms and which model is the best between the two algorithms for classifying hoax news in Indonesian.
In previous studies, research conducted by Lalu Gias, Adiwijaya, and Untari Novia [2] in classifying Indonesian
news using mutual information and support vector machines, it was found that the prediction results with SVM were
the most accurate with an accuracy value of 94.24%. In research on the classification of hoax news using the K-
Nearest Neighbor method and weighting using TF-IDF on the news of the presidential election, the results of the
accuracy value are quite accurate [3], using cosine similarity calculations and validation on all data and then
weighting with TF-IDF, the method K-Nearest Neighbor has an accuracy of 92.30%.
METHODOLOGY
In this study, the researcher uses four categories of datasets, after the data is collected, the data will be preprocessed
before being classified using the Support Vector Machine and K-Nearest Neighbor algorithms.
FIGURE 1 Experiment Design
DATA COLLECTION
The dataset used in this study consists of four categories of data. The first category is scrapping from the website
turnbackhoax.id. data are taken from 2017 – 2019. Scrapping data of 2040 data resulted in 1545 hoax news labels
and 495 fact-labeled news data, where hoax and fact data did not have the same topic and the spread of hoax and
fact data was not balanced. The second category is the modification of the first category by reducing the number of
news labeled hoaxes so that the amount of hoax data and facts is balanced. The data for the second category is 495
factual data and 518 hoax data, where hoax news data and facts do not have the same topic. The third category data
090001-2
is modified data from the first category with the addition of news labeled facts taken from other news sources. The
data for the third category is 2,980 data with 1545 hoaxes and 1435 facts, where hoax news and facts data do not
have the same topic. The fourth category of data is mirroring data, namely related news data (between hoax news
and facts that have the same topic). Hoax data on mirroring data is taken from data scrapping as much as 600 news,
from 600 hoax data researchers are looking for related fact data, where the fact data refers to reference links that
contain narratives on each news.
DATA PREPROCESSING
In the next stage after data collection is complete, the data will be preprocessed. Data preprocessing is carried
out in several stages, namely text cleaning, which is a process to remove noise in the form of characters such as
punctuation marks, symbols, and unknown characters. Case folding is the stage of converting text into a standard
form, which only accepts lowercase letters. Stopword removal (filtering) is the stage of taking important words from
the token results and removing non-descriptive words based on the stopword data list that has been provided. And
the last step is word weighting with TF-IDF.
TF-IDF
Term Frequency – Inversed Document Frequency (TF - IDF) is a feature selection to perform weighting when
the document or data has been extracted. The TF-IDF process begins with weight calculations by integrating TF and
IDF. This step in TF-IDF is to find the number of words that are known (TF) when they have been multiplied by the
number of data/documents in which the word appears (IDF). TF-IDF on SVM is used so that data can be analyzed.
TF-IDF stage is carried out after the data preprocessing process is complete. In the TF-IDF weighting method, the
term weighting of a news item will be carried out by multiplying the Term Frequency value with the Inverse
Document Frequency and the TF-IDF value will be obtained from each word in the news and by adding up the word
weights. Table 1 is an example of word weighting in data using TF-IDF.
CLASSIFICATION ALGORITHM
1. Support Vector Machine Algorithm
The way the SVM algorithm works is to look for a hyperplane with the largest or maximum margin so as not to
misclassify the new data later [5]. The classification process using the SVM method is carried out after the text pre-
processing process and weighting using TF-IDF. news data classification using the SVM method is divided into two
processes, namely the SVM training process and the SVM classifier. In the SVM training process, the aim is to
obtain vector values D, w value and, b value (hyperplane). After the values of w and b (hyperplane) are known, then
the next step is the SVM classifier process, this stage aims to determine whether the news is included in the fact
class or hoax class by utilizing the values of w and b. For testing, when the test value is greater than the hyperplane
it will be included in the hoax news and if it is smaller than the hyperplane value it will be included in the non-hoax
(fact) class. The hyperplane function is obtained from the following mathematical equation [1].
ࢌሺ࢞ሻ ൌ ࢙࢏ࢍ࢔ሺࢃࢀ ࢄ࢏ ൅ ࢈ ൌ ૙ሻ

W = weight of hyperplane position on normal plane
X = vector of input data
b = bias of the plane position relative to the coordinate center
In news classification with the SVM algorithm, the preprocessed data will be divided into training data and test
data with data comparison ratios of 90: 10, 80: 20, and 70: 30, this comparison is to find the most suitable ratio in
the classification to get the best accuracy. The results of the training data are used as a model for test data in the next
process. in this study, for the SVM algorithm using a linear kernel and parameter C with a value of c = 0.001, c = 1
and c = 1000 this is done to find the best c to get the best classification model and also see how the effect of c on the
accuracy value obtained. parameter C is a hyperparameter in the SVM algorithm to control errors when classifying
data. To find out how much C is the best, it can be seen from experimenting with several values and choosing the
value that gives the lowest level of misclassification.
090001-3
TABLE 1 Weighting with TF-IDF
Kata tf.idf
Berita 1 Berita 2 df log(n/df)
Berita 1 Berita 2
menteri 1 0 1 0.84 0.84 0
luar 1 0 1 0.84 0.84 0
negeri 1 0 1 0.84 0.84 0
klarifikasi 1 0 1 0.84 0.84 0
ada 1 0 1 0.84 0.84 0
anak 1 0 1 0.84 0.84 0
wni 1 0 1 0.84 0.84 0
tahan 1 0 1 0.84 0.84 0
teman 0 1 1 0.84 0 0.84
ahok 0 1 1 0.84 0 0.84
bunuh 0 1 1 0.84 0 0.84
diri 0 1 1 0.84 0 0.84
kalah 0 1 1 0.84 0 0.84
pilkada 0 1 1 0.84 0 0.84
weight 6.72 5.04
2. K-Nearest Neighbor Algorithm

The way the KNN algorithm works is by grouping new data with the closest k value distance[6]. The
classification process with the KNN method is carried out after the text pre-processing process and weighting using
TF-IDF. Each data weight or distance will be calculated using the cosine distance which determines the distance
between two points, namely training data and test data. After the distance is known, the value of k will be
determined which will determine if the new data is a fact or a hoax. In news classification with the K-Nearest
Neighbor algorithm, the preprocessed data will be divided into training data and test data with data comparison
ratios of 90: 10, 80: 20, and 70: 30, this comparison is to find the most suitable ratio in the classification to get the
best accuracy. KNN uses the Euclidean formula which can be seen in the following equation [6].
࢔
ࡰሺ࢞ǡ ࢟ሻ ൌ ඨ෍ ሺ ࢞ ࢑ ൅ ࢟࢑ ሻ ૛
࢑ି૚
x = training data
y = test data
In this study, the classification with the KNN algorithm uses Euclidean distance metrics and parameter n as the
value used in the Euclidean metric. The value of n used in this study is Neighbor with a value of n = 3 and n = 5. To
get the best n can be chosen by parameter optimization, for example by doing several experiments with different
values of n. Then evaluate each n value, either by using a confusion matrix or cross-validation, the value of n with
the highest accuracy is the best value for classifying.
RESULTS AND DISCUSSION

In the results and discussion chapter, the evaluation results of the Support Vector Machine and K-Nearest Neighbor
algorithms will be displayed for each data category. A comparison of the level of accuracy and f-measure of the two
algorithms for each category of data visually can be seen in the following diagram.
090001-4
a) b)
c) d)
with:
a) Comparison of the level of accuracy and f-measure in the first category dataset
b) Comparison of the level of accuracy and f-measure in the second category dataset
c) Comparison of the level of accuracy and f-measure in the third category dataset
d) Comparison of the level of accuracy and f-measure in the third category dataset
From the four diagrams above, the researcher will display the evaluation results of the two algorithms by taking the
highest accuracy results from each data category. The retrieval of the highest accuracy results is intended to compare
the SVM and KNN classifications on the best performance obtained in each data category. The evaluation results of
the highest accuracy are shown in the table below.
TABLE 2 The highest accuracy and f-measure results for each dataset category on SVM
SVM Algorithm
Category 1 Category 2 Category 3 Category 4
Accuracy 80% 75% 84% 75%
F-measure 72% 75% 84% 75%
TABLE 3 The highest accuracy and f-measure results for each dataset category on KNN
KNN Algorithm
Category 1 Category 2 Category 3 Category 4

Accuracy 71% 69% 75.2% 64%
F-measure 57% 69% 75% 64%
Based on the table and figure it can be seen that,
Classification of hoax news data in Indonesian for the first category dataset, the SVM algorithm has the highest
accuracy value of 80% and f-measure 72%, this accuracy is obtained at a comparison ratio of 90:10. In the KNN
algorithm for the first category dataset, the accuracy value is 71% and f-measure is 57%, this accuracy is obtained at a
ratio of 70:30. Classification of hoax news data in Indonesian for the second category dataset, in the SVM algorithm
the highest accuracy value is 75% and f-measure 75%, this accuracy is obtained at a comparison ratio of 90:10. in the
KNN algorithm, the highest accuracy value is 69% and f-measure is 69%, this accuracy is obtained at a ratio of 80:20.
Classification of hoax news data in Indonesian for the third category dataset, the SVM algorithm has the highest
accuracy value of 84% and f-measure 84%, this accuracy is obtained at a comparison ratio of 90:10. In the KNN
algorithm, the highest accuracy value is 75.2% and the f-measure is 75%, this accuracy is obtained at a ratio of 70:30.
Classification of hoax news data in Indonesian for the fourth category dataset, in the SVM algorithm the highest
accuracy value is 75% and f-measure 75%, this accuracy is obtained at a ratio of 80:20. In the KNN algorithm, the
accuracy value is 64% and the f-measure is 64%, this accuracy is obtained at a ratio of 90:10.
Classification of news data with the SVM algorithm gets higher accuracy and f-measure values in each data
category compared to the KNN algorithm model, and the best model obtained by the SVM algorithm is in the third
090001-5
category dataset with a dataset comparison of 90: 10 with an accuracy value of 84 % and has an f-measure of 84%.
Based on the implementation results, it can be concluded that the performance evaluation of the Support Vector
Machine algorithm in classifying hoax news data in Indonesian is better than the KNN algorithm, and it is proven that
the Support Vector Machine is a reliable algorithm in solving data classification problems. learning that has been
made by researchers in making learning models, where several stages are passed such as split datasets,
hyperparameters, model evaluations where each process in that stage each algorithm has its own unique features. And
also if we look at the research that has been done with the Support Vector Machine and KNN algorithms, it shows the
same thing where the Support Vector Machine produces the highest accuracy value (research on spam classification
on Youtube using the Naive Bayes, SVM, and KNN methods [4] and on research on online news classification using
SVM and KNN methods) [5].
CONCLUSION
Based on the results of the analysis and discussion of the classification of hoax news, it was concluded that the
best classification model for hoax news or facts was using the Support Vector Machine algorithm, namely the model
with a training and testing data comparison ratio of 90:10, parameter C = 1 and using a linear kernel and the dataset
category is the third category. The accuracy value is 84% and the f-measure is 84%. when the value of c = 1, it
means that the hyperparameter in the SVM algorithm to control errors when classifying data, gets the best results
when the value is 1 or in other words when the value of c = 1, the classification error rate is low.
As for the best classification model for hoax news or facts using the KNN algorithm, namely the model with a
comparison ratio of training data and testing data of 70:30 with n = 3, metrics euclidean distance, and the dataset
category, namely the third category. The accuracy value is 75.2% and the f-measure is 75%. When n = 3 which
means when the value used in the Euclidean metric is 3, the best classification model for KNN is obtained, where at
the time of classification many are correct according to the actual label of the news.
So between the two algorithms, the best model is obtained by using the classification model of the Support
Vector Machine algorithm. And it can also be concluded that the distribution ratio and the amount of data affect the
accuracy of the model created. It can be seen from the accuracy results obtained in each training data distribution
ratio and the testing data distribution is different.
SUGGESTION
Based on the results of the analysis and conclusions, to get better classification, further research can use larger data
provided that the hoax label data and facts are balanced and use scrapping on other news media to get datasets in
real-time.
ACKNOWLEDGMENTS
The author is very grateful to Research and Community Services Unit (LPPM) Institut Teknologi Del for supporting
this publication.
BIBLIOGRAPHY
1. E. Rasywir and A. Purwarianti, "Experiments on the Machine Learning-Based Indonesian Hoax News
Classification System," J. Cybermatics, vol. 3, no. 2, pp. 1–8, 2015, [Online]. Available:
https://www.mendeley.com/import/.
2. LG Irham, A. Adiwijaya, and UN Wisesty, “Classification of Indonesian News Using Mutual Information
and Support Vector Machines,” J. Media Inform. Budidharma, vol. 3, no. 4, p. 284, 2019,
doi:10.30865/mib.v3i4.1410.
3. FN Rozi and DH Sulistyawati, “CLASSIFICATION OF ELECTION HOAX NEWS USING MODIFIED
K-NEAREST NEIGHBOR METHOD AND WEIGHTING USING TF-IDF,” J. Sist. Computing., vol. 10,
no. 2, pp. 49–54, 2019.
4. Burhanudin, Y. Musa'adah, and Y. Wihardi, "Classification of Spam Comments on Youtube Using the
Naïve Bayes Method, Support Vector Machine, and K-Nearest Neighbors," J. Inform. and Computing., vol.
3, no. 2, pp. 54–59, 2018.
5. SN Asiyah and K. Fithriasari, "Online News Classification Using Support Vector Machine and K-Nearest
Neighbor Online News Classification Method Using Support Vector Machine and K-Nearest," J. Science
and Arts ITS, vol. 5, no. 2, 2016.
6. Pinandito, A., Perdana, R. S., Saputra, M. C., & Az-Zahra, H. M. (2017). Spam detection framework for
Android Twitter application using Naive Bayes and K-Nearest Neighbor classifiers. ACM International
Conference Proceeding Series, 77–82. https://doi.org/10.1145/3056662.3056704
090001-6
Sentiment Analysis of Government Regulations Regarding
the Implementation of Reverse-Transcriptase Polymerase
Chain Reaction (RT-PCR) During the Covid-19 Pandemic in
Indonesia (Case Study: Air Transportation Mode)
A.Prisdayanti1,a , I.Budi1, A.B.Santoso1, and P.K.Putra1
1
Faculty of Computer Science,
Universitas Indonesia, Jakarta 10430, Indonesia
a)
Corresponding author: anella.prisdayanti@ui.ac.id
Abstract. Many people express their opinions regarding policies or rules imposed by the government through Twitter
social media. This study was conducted to determine public sentiment regarding the rules for implementing the reverse-
transcriptase polymerase chain reaction (RT-PCR) test as one of the requirements for using air transportation. The research
was conducted using the Random Forest and K-Nearest Neighbor (KNN) algorithms and the CRISP-DM research method.
Twitter data used is classified into negative, neutral and positive. The results of the sentiment analysis showed that the
KNN algorithm outperformed Random Forest with an accuracy value of 75.63% and the sampling technique used was
stratified sampling. The results of the sentiment analysis show that the public has a negative sentiment towards the policies
issued by the government.
INTRODUCTION
At the beginning of 2020 the world was shocked by the emergence of a virus outbreak known as Novel Coronavirus
or higher called as Covid-19. This virus is thought to have originated in the city of Wuhan, China. Corona virus is one
type of virus that has a high rate of spread. So on March 11, 2020, the World Health Organization (WHO) declared
that the corona virus had come to be a global pandemic. Based on a report from the WHO, globally the entire range
of instances of the unfold of the corona virus is 221,134,742 cases with 4,574,089 deaths 1.
In Indonesia, on March 2, 2020, for the first time, 2 residents were detected who were positive for the Corona virus
in Depok, West Java 2. Since then, based on data from the Covid-19 task force records, cases of confirmed positive
cases of the corona virus have been increasing every day. The wider spread of the virus has an impact not only in the
health sector but also in the economy, education, and others.
The high number of cases of the spread of Covid-19 in Indonesia has forced the government to act to overcome
this. One of the government's efforts to break the chain of the spread of Covid-19 is to implement an emergency PPKM
(Enforcement of Community Restrictions) starting on July 3, 2021 3. During the emergency PPKM period, the
government issued regulations on transportation to reduce community mobility. One of the areas affected by the
regulation is air transportation. For air transportation, the government enforces a rule that passengers must have a
negative certificate of reverse-transcriptase polymerase chain reaction (RT-PCR) test whose samples are taken within
a maximum period of 2x24 hours. This rule has been implemented since July 5, 2021 4.
The rules imposed by the government reap many pros and cons from the community. Some agree and some
disagree with the implementation of the PCR test as one of the requirements for using air transportation. Many people
have expressed their opinion regarding the regulation through social media. One of the social media that is widely
used by the public to express opinions is through Twitter. Twitter is a very popular microblogging site where users
are allowed to provide descriptions of ideas and opinions up to 140 characters 5. Twitter is widely used as a medium
for sharing information and reporting news 6. Based on this, Twitter can be the right source of data for sentiment
analysis.
Sentiment analysis is an activity carried out to analyze a person's opinion, assessment, or attitude towards a service,
product, or issue. Sentiment analysis is good to apply because the analysis carried out is based on the opinions and
views felt by someone. The opinions of others are often used by someone to make a decision 7.
090002-1
In this study, sentiment analysis was carried out to find out public opinion regarding government policies regarding
the implementation of PCR swabs as one of the requirements for using air transportation modes. The opinion is
obtained from Tweet data submitted by the public. The data is then classified into positive, negative or neutral for
further analysis.
The algorithm used in this research is K-Nearest Neighbor (KNN). KNN is an algorithm that is widely used to
perform sentiment analysis. KNN is a classifier algorithm that performs object classification based on learning from
the data closest to the object. This method is also known to be simple and easy to implement 8. Furthermore, the
performance of KNN will be compared with the ensemble method, namely Random Forest. Random Forest is one of
the popular and widely used ensemble methods because it is easy to learn and use, even for the layman. In addition,
random forest has also been widely applied in various fields such as remote sensing, credit rating analysis and others
9
.
MATERIALS
K-Nearest Neighbor (KNN)
Based on research conducted by 10, it shows that sentiment analysis of the 2013 curriculum using the KNN method
gets an accuracy value of 96% with a value of k=5. Another study conducted by 11, shows that the accuracy value using
KNN gets higher results when compared to Term Graph and Naïve Bayes algorithm. The average accuracy value
obtained using the KNN method is 98.95%, 98.72% for the Term Graph and 62.66% for the Naïve Bayes method.
K-Nearest Neighbor (KNN) is a machine learning algorithm that classifies objects based on the closest data learning
to the object. The KNN algorithm is carried out by looking for the value of k in the training data with the closest distance
that has similarities to the objects contained in the test data 12. The process of classifying objects is done using the
Euclidean equation. Euclidean is used to determine the distance between the training object and the test object 13.
Here is how the KNN algorithm works 14 :
x Calculate the similarity of all data representatives in the M model of the data set (dx) to be classified.
x If dx only consists of 1 classification then Class (dy), Similarity (dy), Number (dy), Representative (dy),
Euclidean distance between dx and dy < Similarity (dy), where dx is classified as dy
x If dx is categorized by at least 2 different classes, then dt is classified as the representative category with the
largest dy and the environment will include the largest data set into the training dataset.
x If there is no data representation in the M model that includes dx, then classify dx whose boundary is closest to
dm.
Random Forest
Random forest is an ensemble method of the decision tree algorithm where the predicted value displayed is a
combination of each tree in the ensemble. Each tree of the decision tree that is built is formed using a random data set
from training data with a fixed data spread probability.
How the Random Forest algorithm works 15:
x Select X record based on existing dataset. Where X is the number of trees built
x The number of decision trees is built based on the specified number of X
x Build a decision tree according to the existing dataset.
x For regression, each tree of the forest predicts a Y value for the new record. The final value is calculated based
on the average of all values predicted by each tree of the forest.
x For classification, each tree of the forest will predict a new record to fall into the specified category.
METHODOLOGY
The sentiment analysis process is carried out through several stages. This step is carried out by adopting the CRISP-
DM (Cross-Industry Standard Process for Data Mining) methodology. CRISP-DM is a data mining model that is
widely used as a standard for data processing in data mining 16. The CRISP-DM model also aims to make data
processing more reliable, more manageable, repeatable and faster 17. The stages of research using CRISP-DM can be
seen in figure 1.
090002-2
FIGURE 1. Data Analysis Stages
Business Understanding
This study aims to determine public opinion regarding government policies in making Negative PCR Swab rules
taken within 2x24 hours as one of the requirements to be able to use air transportation modes. The research question
from this research is "How is the public sentiment regarding the PCR Swab rule as one of the requirements for using
air transportation mode?". Another research question is "How is sentiment analysis using K-Nearest Neighbor
(KNN)?".
Data Understanding
This stage begins with data collection and data processing to identify data quality problems and find previously
unknown information from the data 17. In this study, data collection was carried out by pulling Twitter data containing
opinions related to the PCR swab policy for air transportation modes. The data collected is Indonesian tweet data
containing the words "swab, PCR and pesawat". The data was taken from July 5, 2021-August 31, 2021. The data that
has been collected is then stored in a csv file for later processing.
Data Preparation
This stage is a stage that includes activities to build a dataset. At this stage, the appropriate attribute selection, data
cleaning and data transformation are carried out so that the data can be used as a dataset to build a model 17.
1. Cleaning
At this stage, the data that has been collected is then cleaned. The cleaning is carried out by removing symbols,
hashtags, mentions, and urls contained in tweets to produce better data and remove noise from the data.
2. Data Transformation
At this stage, data transformation processes such as case folding and filtering are carried out. Case folding is
a step taken to convert data that has capital letters into lowercase letters. Filtering is a step taken to get rid of
words that have less meaning.
Modelling
At this stage, the model is built from the data that has been prepared. This study uses the KNN and Random Forest
algorithms to perform sentiment analysis and Cross Validation is used to evaluate the performance of the model built
Evaluation
After building the model, the next step is to evaluate the performance of the built model. At this stage, trials of the
model built using data testing are carried out and see how well the model is built to identify sentiment.
Deployment
The last stage is to make a report and make a summary related to the activities of the analysis process carried out.
At this stage, conclusions are drawn regarding public opinion on policies issued by the government regarding PCR
swabs as a condition for using air transportation modes.
RESULTS
Business Understanding
At this stage, data is collected from Twitter regarding public opinion regarding the rules issued by the government
regarding PCR swabs as one of the requirements for using air transportation modes. The data is taken using the tweepy
090002-3
library which is a package contained in the python programming language. The collected fields are id, created at, and
tweet and then saved in csv format. The data taken is a tweet containing the query "swab pcr pesawat" and in
Indonesian.
Data Understanding
To better understand the data, labeling of the previously collected data was carried out. Labeling is done into 3
classes, namely positive, negative, and neutral which can be seen in table 1. The amount of data collected is 395 data
with the distribution of data, namely negative 246 tweets, positive 120 tweets and neutral 29 tweets.
TABLE 1. Sample Data Labelling of Tweet

Tweet Sentiment
Yang paling rawan adalah pesawat ada di ruang tertutup selama jangka waktu tertentu, sama seperti kabin bis AC, Positif
ini rawan penularan. PCR sebelum naik pesawat menurut saya tetap perlu.
(The most vulnerable is that the plane is in an enclosed space for a certain period of time, just like the cabin of an
air-conditioned bus, this is prone to transmission. PCR before boarding the plane in my opinion is still necessary.)
Coba kita yang sudah di vaksin, kalo naik pesawat cukup tunjuk kartu vaksin, terus bebas penerbangan, ini mah Negatif
disuruh tes PCR lagi, jadi pusing
(Try those of us who have been vaccinated, if you get on a plane, you just need to show your vaccine card, then
you're free to fly. I'm told to do a PCR test again, so I'm dizzy.)
tolong salah satu ada yang bisa bantu mempertegas penjelasan apabila naik pesawat dari atau ke pulau jawa-bali Netral
apabila sudah vaksin kedua apakah harus tes pcr atau cukup antigen?
(Can someone please help clarify the explanation if you take a plane from or to the island of Java-Bali, if you have
the second vaccine, do you have to do a PCR test or do you have enough antigen?)
Data Preparation
The steps taken in the process of preparing the data can be seen in table 2.
TABLE 2. Data Preparation Process

Process Data
Real Data Kalau harga masih tinggi terus, apa ini mau menanggulangi Covid-19 atau berbisnis ditengah
pandemi? Dan kalau bisa persyaratan untuk perjalanan melalui pesawat terbang enggak usah PCR,
cukup dengan swab antigen.karena semua penumpang sudah prokes.
(If prices are still high, is this going to deal with Covid-19 or doing business in the middle of a
pandemic? And if possible, the requirements for traveling by airplane don't need PCR, it's enough
with an antigen swab. Because all passengers have been prokes)
Cleaning Kalau harga masih tinggi terus apa ini mau menanggulangi Covid atau berbisnis ditengah pandemi
Dan kalau bisa persyaratan untuk perjalanan melalui pesawat terbang enggak usah PCR cukup
dengan swab antigen karena semua penumpang sudah prokes
(If prices are still high is this going to deal with Covid or doing business in the middle of a pandemic
And if possible the requirements for traveling by airplane don't need PCR it's enough with an antigen
swab Because all passengers have been prokes)
Case Folding kalau harga masih tinggi terus apa ini mau menanggulangi covid atau berbisnis ditengah pandemi
dan kalau bisa persyaratan untuk perjalanan melalui pesawat terbang enggak usah pcr cukup dengan
swab antigen karena semua penumpang sudah prokes
(if prices are still high is this going to deal with covid or doing business in the middle of a pandemic
and if possible the requirements for traveling by airplane don't need pcr it's enough with an antigen
swab because all passengers have been prokes)
Stemming kalau harga masih tinggi menanggulangi covid bisnis tengah pandemi kalau persyaratan perjalanan
melalui pesawat terbang enggak usah pcr cukup swab antigen semua penumpang sudah prokes
(if the price is still high dealing with covid, business is in the middle of a pandemic if the
requirements for travel by airplane are no need for pcr just swab antigen all passengers have been
prokes)
Filtering harga tinggi tanggulangi covid bisnis tengah pandemi syarat jalan pesawat terbang pcr swab antigen
penumpang prokes
(high price to deal with covid, business is in the middle of a pandemic airplane road conditions pcr
antigen swab passenger prokes)
Tokenize “harga”, “tinggi”, “tanggulangi”, “covid”, “bisnis”, “tengah”, “pandemi”, “syarat”, “jalan”,
“pesawat”, “terbang”, “pcr”, “swab”, “antigen”, “penumpang”, “prokes”
090002-4
(“prices”, “high”, “to deal”, “covid”, “business”, “middle”, “pandemic”, “requirement”, “travel”,
“air”, “plane”, “pcr”, “swab”, “antigen”, “passenggers”, “prokes”)
Testing and Evaluation

1. Split of Data Testing and Training
After going through the data preparation process, then the distribution for testing and training data is carried out.
The process of data dividing is carried out using k-fold cross validation. K-fold cross validation is one of the techniques
used to evaluate the performance of a method as a whole by grouping data as much as the specified k value and
dividing it randomly 18.
In this study, the k values used are 5, 10, 15, and 20 fold. The dataset will be divided as much as the fold value,
where one part will be used as test data, the rest of the other group will be used as training data. The data used as
training data is then used by the machine as material for the learning process. The test data will be used as a data set
to test the performance of the final model.
2. Analysis and Evaluation of Algorithm

The evaluation of the KNN and Random Forest algorithms is carried out using k-fold cross validation and can be
seen in table 3. The k-fold evaluation process was carried out using shuffeld sampling and stratified sampling. Shuffled
sampling is a technique of sampling which is done by taking data from training data randomly where each element
has the same selection probability. Stratified sampling is a sampling technique which is done by ensuring that each
group in a population is adequately represented in the sample 19.
Based on table 3, it can be seen that the best accuracy is obtained by the KNN algorithm with a stratified sampling
technique and the value of fold = 20 (75.63%). In addition, it can also be seen that both the KNN and Random Forest
algorithms have the best accuracy obtained by using a stratified sampling technique. It also proves that stratified
sampling is a good sampling technique to apply. This is because this technique is believed to be able to choose a better
sample that represents the distribution of data from the entire dataset 20.
TABLE 3. Accuracy Results

Accuracy
K-Fold Shuffled Sampling Stratified Sampling
KNN RF KNN RF
5 72.58% 61.41% 72.34% 62.70%
10 74.35% 63.68% 73.58% 61.94%
15 74.65% 62.43% 74.88% 61.69%
20 74.87% 63.97% 75.63% 64.16%
The model evaluation process is also carried out by measuring recall and precision. The confusion matrix can be
seen in table 4.
TABLE 4. Confussion Matrix
K-Nearest Neighbor (Accuracy = 75.63%)
True Negatif True Positif True Netral Precision
Pred.Negatif 212 39 16 79.40%
Pred.Positif 31 81 7 68.07%
Pred.Netral 3 0 5 62.50%
Class recal 86.18% 67.50% 17.86%
Random Forest (Accuracy = 64.16%)
True Negatif True Positif True Netral Precision
Pred.Negatif 239 106 27 64.25%
Pred.Positif 6 14 1 66.67%
Pred.Netral 1 0 0 0.00%
Class Recall 97.15% 11.67% 0.00%
Deployment : Result of Analysis Sentiment

The final step after evaluating the model is analyzing the classification results obtained. Based on table 4, for the
KNN algorithm which has the highest accuracy, it can be seen that the highest precision and recall values are obtained
by negative sentiment with a precision value of 79.40% and a recall of 86.18%. This shows that the system can
090002-5
recognize the features of data that are labeled negatively. While the lowest precision and recall values were obtained
by neutral sentiment with a precision value of 62.50% and 17.86% recall. This is because neutral sentiment has a
confusing feature that is difficult for the system to recognize. Meanwhile, for the Random Forest algorithm, it can be
seen that the algorithm can recognize the features of the negative class well but it is still difficult to recognize the
features of the positive and neutral classes.
In addition, based on table 4, it can also be seen that based on the sentiment analysis process carried out with 20-
folds, the KNN algorithm can recognize as many as 212 negative comments, 81 positive comments and 5 neutral
comments. The Random Forest algorithm can recognize 239 negative comments, 14 positive comments and 0 neutral
comments.
CONCLUSION
Evaluation of sentiment analysis performed using the K-Nearest Neighbor and Random Forest algorithms was
carried out by testing using the values of accuracy, precision and recall. Based on the tests carried out, KNN
outperformed Random Forest with an accuracy value of 75.63%. In addition, the highest accuracy value, either KNN
or Random Forest, was obtained using a stratified sampling technique with k-fold = 20. This indicates that the data
sampling technique affects the resulting accuracy value. In addition, based on the analysis, it was concluded that the
community had negative sentiments related to government policies that imposed negative PCR swab rules within a
period of 2x24 hours as one of the conditions for using air transportation modes. Some suggestions for further research
are 1) use other classification algorithms to perform sentiment analysis such as Naive Bayes, SVM, and decision trees;
2) use larger data with more even distribution of data; 3) using data obtained from other platforms such as Instagram
and Facebook.
REFERENCES
1
WHO, WHO Coronavirus (COVID-19) Dashboard, (https://covid19.who.int).
2
Indonesia.go.id, Kasus Covid-19 Pertama, Masyarakat Jangan Panik, (https://indonesia.go.id/narasi/indonesia-
dalam-angka/ekonomi/kasus-covid-19-pertama-masyarakat-jangan-panik).
3
covid19.go.id, PPKM Darurat: Tugas Bersama Turunkan Pandemi Covid-19 - Berita Terkini,
(https://covid19.go.id/p/berita/ppkm-darurat-tugas-bersama-turunkan-pandemi-covid-19).
4
dephuh.go.id, Aturan Bertransportasi Dalam Negeri di Masa PPKM Darurat, (http://dephub.go.id/post/read/aturan-
bertransportasi-dalam-negeri-di-masa-ppkm-darurat).
5
M. Rezwanul, A. Ali, and A. Rahman, Int. J. Adv. Comput. Sci. Appl. 8, 19 (2017).
6
A. Java, X. Song, T. Finin, and B. Tseng, Jt. Ninth WebKDD First SNA-KDD 2007 Work. Web Min. Soc. Netw.
Anal. 56 (2007).
7
S. Alhojely, Sentiment Analysis and Opinion Mining: A Survey (2016).
8
A. Desai and C. Shrihari R., Int. J. Comput. Appl. 111, 12 (2015).
9
Y. Al Amrani, M. Lazaar, and K.E. El Kadirp, Procedia Comput. Sci. 127, 511 (2018).
10
M.R. Irfan, M.A. Fauzi, T. Tibyani, and N.D. Mentari, Int. J. Electr. Comput. Eng. 8, 5409 (2018).
11
V. Bijalwan, V. Kumar, P. Kumari, and J. Pascual, Int. J. Database Theory Appl. 7, 61 (2014).
12
R. Agrawal, Int. J. Comput. Appl. 105, 13 (2014).
13
A.S. Thakur and N. Sahayam, Int. J. Emerg. Technol. Adv. Eng. 3, 587 (2013).
14
G. Guo, H. Wang, D. Bell, Y. Bi, and K. Greer, Lect. Notes Comput. Sci. (Including Subser. Lect. Notes Artif.
Intell. Lect. Notes Bioinformatics) 2888, 986 (2003).
15
P. Karthika, R. Murugeswari, and R. Manoranjithem, IEEE Int. Conf. Intell. Tech. Control. Optim. Signal Process.
INCOS 2019 1 (2019).
16
P.C. Ncr, J.C. Spss, R.K. Ncr, T.K. Spss, T.R. Daimlerchrysler, C.S. Spss, and R.W. Daimlerchrysler, SPSS Inc 78,
1 (2000).
17
R. Wirth and J. Hipp, Proc. Fourth Int. Conf. Pract. Appl. Knowl. Discov. Data Min. 29 (2000).
18
A. Rohani, M. Taki, and M. Abdollahpour, Renew. Energy 115, 411 (2018).
19
L. Qian, G. Zhou, F. Kong, and Q. Zhu, EMNLP 2009 - Proc. 2009 Conf. Empir. Methods Nat. Lang. Process. A
Meet. SIGDAT, a Spec. Interes. Gr. ACL, Held Conjunction with ACL-IJCNLP 2009 1437 (2009).
20
S. Li, G. Zhou, Z. Wang, S.Y.M. Lee, and R. Wang, Int. Conf. Inf. Knowl. Manag. Proc. 2469 (2011).
090002-6
Abortion Drug Products Classification Using Text Mining:
A Case Study of PT XYZ
Muhammad Faisal Mazidnianto1,a), Anella Prisdayanti Damanik1,b), Indra Budi1,c),
Aris Budi Santoso1,d), Prabu Kresna Putra1,e)
Author Affiliation
1
Faculty of Computer Science
Universitas Indonesia
Jakarta, Indonesia
Author Emails
a)
Corresponding author: muhammad.faisal03@ui.ac.id
b)
anella.prisdayanti@ui.ac.id
c)
indra@cs.ui.ac.id
d)
aris.budi@ui.ac.id
e)
prabu.kresna@ui.ac.id
Abstract. There are several online media closed by the Ministry of Communication and Informatics for selling abortion
drug. To keep PT XYZ from being shut down by the Ministry of Communication and Information, PT XYZ pays
attention for the circulation of this abortion drug by developing the pending system. However, the pending system only
waited for the title of the product using specific keywords related to the drug input by the team so that there were still
abortion drug products that passed from the system because there were products that used general keywords and
sometime seller play with the keywords. Therefore, this study is conducted to build text classification model derived from
the existing abortion drug products in PT XYZ which will be used for the detection of future abortion drug. This study
uses the CRISP-DM model for the data mining life cycle. This study compares the model with price features and without
price features. The best result is Naive Bayes model with price features generates 99,34% of accuracy and 99,34% of f1-
score. Keywords— Text Mining, Classification, Abortion Drug, CRISP-DM
INTRODUCTION
In 2015 Ministry of Communication and Information Technology in Indonesia closed 300 thousand websites for
selling illegal drug and it was stated that out of all the illegal drug sellers, there were more abortion or abortion drug
sellers. Finally, in 2017-2019, the Ministry of Communication and Information closed 96 sites that openly sell
abortion drug1. Closing the site that sells abortion drug is based on the Indonesia Criminal Code Law in Article 346
which states that a woman who deliberately aborted or killed her womb or ordered someone else to do so, can be
punished with a maximum imprisonment of four years. Besides, it is also clarified in the Criminal Code Law in
article 349 which states that if a doctor, midwife, or medicine worker helps to commit a crime under Article 346,
then the penalties specified in that article can be added by one third.
In an interview with one of the PT XYZ Risk Management Team, it was stated that in handling abortion drug
products to be bought and sold on the PT XYZ application, the Risk Management Team already has a Pending
System that detects abortion drug products based on certain keywords making the store enter into a surveillance
status if it sells abortion drug so that the product cannot be purchased by the buyer from the PT XYZ application.
The system only detects through the product name. After that, if the product is an abortion drug product, it will be
removed from the PT XYZ application. Through the system, PT XYZ hoped that there is no opportunity for these
products to be purchased by buyers from the PT XYZ application.
After the interview, there are still some abortion drug products in PT XYZ application whose stores were not
under surveillance status so there is still a possibility that buyers can buy from the PT XYZ application which can be
090003-1
seen in FIGURE 1, there are still abortion drug products that were circulating even two weeks from the day where
the authors took the data. This indicates that the reality is far from expectations.
FIGURE 1. One Hundred Abortion Drug Products that have been Circulating in PT XYZ Application
Even though there is Pending System, the Risk Management Team is not completely hands-off, they also
perform manual sweeping to anticipate products that pass the pending system by performing queries directly into the
database. They do the manual sweeping because sometime abortion drug products contain only general keyword in
their product name and also seller often try to change the product name that cannot be found by the pending system.
Besides, it is also stated that sometimes the query process is quite long because the product data in PT XYZ is quite
a lot. The long process allows the seller to have the opportunity to sell his/her abortion drug products from the PT
XYZ application so that expectations are not achieved.
Based on the identification results, the root of the problem that will be raised for this study is the absence of
machine learning to handle the circulation of abortion drug. By using machine learning, the abortion drug products
can be detected by learning the pattern not only through the product name but also product description or others data
related to the products. Choosing the right technique in text mining on machine learning can reduce time and effort
to find relevant patterns for analysis and decision making2. Based on this, the following research questions were
obtained "How can the appropriate text mining classification model be used to detect abortion drug products to
make it easier and faster for the detection process?".
LITERATURE REVIEW
A. Abortion Drug
One of the ways to perform this abortion is to use drug. In the interview with pharmacist, there is actually no
drug or pill that is intended for abortion, but there is drug that are intended to cure other diseases but are misused for
the purpose of abortion. For example, such as Misoprotosol, it is actually used for peptic ulcer disease, but has
abnormal side effects on the fetus or risk to the fetus for pregnant women. It is explained by The National of Drug
and Food Control (BPOM) in its official website that the drug offered online media as abortion drug are Cytotec and
Gastrul is a drug with the active substance Misoprotosol and registered in BPOM with indications as peptic ulcers3.
B. Text Mining
Text mining is the discovery of some previously unknown information by extracting that information from a
large and unstructured set of text 4. Text Mining is part of data mining that specializes in text data 5. By using text
mining, company can dig deeper information from the collection of text that is then processed and become the
information needed by the company to help decision making.
C. Preprocessing
Preprocessing supports text mining efforts to take advantage of the many different elements contained in natural
language documents to convert them from unstructured and weakly structured text data to structured data 6. The
preprocessing process is shown in FIGURE 2.
FIGURE 2. Preprocessing in Text Mining
090003-2
Transforming cases is the stage of converting all characters (upper and lower case) in the text to lowercase. This
stage is used to eliminate homologous words that only differ by the upper and lower case. Tokenizing is the stage
carried out in text mining which includes words or sentences that must be separated word by word to improve
processing. So, all words are separated in a sentence, and all punctuation marks are discarded because they cannot
represent any group. This simplifies the calculations in the next step. Filtering Stopwords is a stage used to remove
unnecessary things such as words that appear too often or too rarely in sentences or text documents. It is also used to
remove unnecessary words or words that have no special meaning such as conjunctions or conjunctions which are
applied to reduce processing time or computational complexity. Featured Extraction based on TF-IDF, which is one
of the most well-known algorithms used in text mining research. Term Frequency (TF) is the number of times a
word is repeated in a document and Inverse Document Frequency (IDF) is an algorithm used to calculate the inverse
probability of finding a word in the text.
D. Classification Model
From data that has been converted into structured data that has been labeled, it will be divided into training data
and testing data. The method for doing this division is k-fold cross validation. In this process, training data is used to
create a classification model which will be used to label an input data that you want to find out. The classification
process builds a predictive model that can classify documents automatically6. The following are the algorithms and
techniques that is used in this study:
1. Naive Bayes (NB)
Naive Bayes is an algorithm of a supervised learning algorithm based on the Bayes theorem which has
been widely used because of its simplicity and reasonable performance7. Naive Bayes classification is used
as a probabilistic classifier. Naive Bayes gave us very good results when used to analyze text data 8. The
feature extraction that will be used in this study uses tf-idf in the form of numerical data, so Multinomial
Naive Bayes is the right classification model9.
2. Support Vector Machine (SVM)
Support Vector Machine (SVM) is a regression and classification technique that requires documents
previously classified as training data (supervised learning)6. SVM aims to obtain a separator function
(hyperplane) that separates data by maximizing the distance between positive class data and the closest data
in the negative class which can be seen in FIGURE 3.
FIGURE 3. SVM Hyperlanes

3. k-fold cross validation
Cross validation is a method used to divide the sample data into several sample partitions. Initially, the
analysis was carried out on the first sample partition then the subsequent sample partitions were used to
confirm and validate the initial analysis. Each sample partition contains training data and testing data which
will be used by the classification model algorithm in testing the performance of the classification model.
From several sample partitions, iteration is carried out from the first to the last sample partition so that the
results of the classification model performance based on each sample partition are averaged10.
090003-3
DISCUSSION AND RESULT
This section will explain the flow of the research to be carried out. This study adopts the CRISP-DM
framework, it can be seen in FIGURE 4.
FIGURE 4. CRISP-DM
A. Business Understanding
This phase focuses on understanding the goals and needs of the project from a business perspective, which then
converts the knowledge into a definition of a problem that can be solved by data mining, and an initial plan designed
to achieve that goal11.
At this stage, identification of the root of the problem is carried out using two methods, those are interview and
observation.
From the interview, based on Article 349 of the Criminal Code Law, PT XYZ has built a detection system that
detects abortion drug based on the product name only. Even though this system already exists, from observations
made after conducting interviews, it was still found that one hundred abortion drug products are listed in the PT
XYZ application, so there is still the possibility that buyers can buy this drug. This is because there is general
keywords in product name so that the system cannot detect the product and there are sellers of the abortion drug who
"play" the product name, for example, one product that is abused as an abortion drug is "cytotec", the seller makes a
title with the name "cyto”, “citoteck”, and so on.
B. Data Understanding
This phase begins with the initial data collection and continues with activities to understand the data so that they
are familiar with the data. Then, identify data quality problems, find the first insights into the data, detect what is
interesting from the hidden information11.
At this stage, the data that is used in this study were collected. Data collection is done by querying the database
from PT XYZ. Total of data used was 10105 data with a comparison of 5007 data for abortion drug product data and
5098 data for non-abortion drug product data. Data that had been detected as abortion drug will be labeled as
abortion drug and for non-abortion drug product data at PT XYZ, this study use product from special type of shop
which seller need to pay for it to get some privileges. One of the privileges is seller get dedicated business team as a
consultant for the shop to grow in PT XYZ application. As a consultant, one of the business team tasks is to ensure
all of the products are not violated products based on term and condition in PT XYZ. The product components used
in building the model are product titles, product descriptions, and product price
C. Data Preparation
This phase includes all activities in compiling the final dataset (data that will be used in building a model) from
the initial raw data. This activity includes selecting the tables, records, and attributes that you want to use. Apart
from that, data cleaning was also carried out, constructing new attributes, and transforming the data into modeling
materials11.
At this stage, data that has been labeled and is still in the form of unstructured data will be processed into
structured data that can be used to build classification models. At this stage, preprocessing steps are carried out,
namely transformation cases, tokenizing, stopword filtering, data cleaning (remove url, symbol, number, emoji, etc),
feature extraction with TF-IDF. The result of this stage is structured data that can be used for the classification
model creation process.
090003-4
D. Modelling
In this phase, various modeling techniques are selected and applied, and the parameters are calibrated to optimal
values11. At this stage, a classification model is formed. Before using data that has been processed, the data is
divided into two, consisting of training data and testing data. Divide the two data using 10-fold cross validation.
The model built in this study uses two classification algorithms, namely Naive Bayes and SVM. Because the
feature extraction used is TF-IDF so that the data is numerical, so for Naive Bayes, this study uses the Multinomial
Naive Bayes algorithm. The SVM used in this study uses a linear kernel type. This study compares model with price
features and without price features which the categorization looks like seen in TABLE 1. The categorization is
obtained from looking at the pattern of the price range of the abortion drug obtained that can be seen in FIGURE 5.
The picture shows that the price range of the abortion drug reaches up to two million rupiah. These models are built
using the pandas, numpy, nltk, nlp_id, and sklearn libraries in the python programming language.
TABLE 1. Price Category
Price Features Specification
low_price price <= Rp. 500.000
low_med_price price > Rp. 500.000 and price <= Rp.
1.000.000
med_price price > Rp. 1.000.000 and price <=
Rp. 1.500.000
med_high_price price > Rp. 1.500.000 and price <=
Rp. 2.000.000
high_price price > Rp. 2.000.000
FIGURE 5. Price Data Distribution

E. Evaluation
After these models are created, an evaluation is carried out to determine which model is the most effective
model for machine learning4. To evaluate classification models, a confusion matrix method can be used12.
To ensure that the model reaches the business goals correctly, this evaluation is done by building a
configuration matrix and making comparisons between several ratios, those are accuracy (ratio expressed as the
proportion of correctly classified cases of all cases13), precision (ratio expressed as the proportion of positive cases
correctly recognized as positive in all cases classified as positive13), recall (ratio expressed as the proportion of
positive cases correctly recognized as positive above all actual positive cases13). f1-score (measurement consisting
of a combination of recall and precision13).
After conducting data preparation, for the formation of training and testing data, this study uses k-fold cross-
validation with a k value of 10. To evaluate the model, this study includes parameters such as accuracy, precision,
recall, and f1-score which can be seen in TABLE 2.
TABLE 2. Model Evaluation
Accuracy Precision Recall f1-score
NB + PF 99,34% 99,51% 99,18% 99,34%
SVM+ PF 99,29% 99,08% 99,51% 99,30%
NB 99,17% 99,51% 98,84% 99,17%
SVM 99,23% 99,02% 99,45% 99,24%
F. Deployment
In this phase, conclude the results of the data mining activities and conclude important information11. At this
stage, from the information obtained such as the result of the evaluation of the classification model of abortion drug,
this study makes some conclusion so that important information which might be useful for further research.
In this case study, the values of false-positive and false-negative have the same importance. If the model has a
high false-positive value, it indicates that the model predicts a product that is not an abortion drug but is considered
an abortion drug. This can make non-abortion drug products can be changed their status into surveillance status so
that buyers cannot buy these products until the seller protests to PT XYZ to be reactivated. This can harm seller
because during the reactivation process, the seller does not get income. If the model has a high false-negative value,
it indicates that the model predicts a product that is an abortion drug but is considered a non-abortion drug product.
This makes PT XYZ violate the Criminal Code Law which has been described previously and allows buyers to be
able to buy the abortion drug product. Because these two values are important, the right evaluation value is to use
the accuracy and f1-score. From the model that has been built, the model that uses the Naive Bayes algorithm using
the price feature has the best accuracy and f1-score value which is 99.34%.
090003-5
In this case, accuracy and f1-score have an important role. Therefore, Naive Bayes model that uses the price
feature is better than the other models. Through this study, it is found that adding features that have patterns can
improve the performance of the model, in this case, the price feature.
CONCLUSION
This study was conducted to build the best model for detecting abortion drug products at PT XYZ which was
guided by relevant previous studies. The data components used in this study are product titles, product descriptions,
and product prices. it was found that the Naive Bayes model using the price feature produced the best results,
namely with an accuracy value of 99.34% and an f1-score value of 99.34%. Based on this study, there are several
suggestions for further research including using larger data to reach a wider variety of product data. In addition, it
may be possible to use several other classification algorithms in building a model for comparison and consider k-
fold cross-validation with a more diverse k.
REFERENCES
1
D. Pramita, (n.d.).
2
R. Talib, M.K. Hanif, S. Ayesha, and F. Fatima, Int. J. Adv. Comput. Sci. Appl. 7, (2016).
3
Badan Pengawas Obat dan Makanan Republik Indonesia, Bpom (2015).
4
S.M.H. Dadgar, M.S. Araghi, and M.M. Farahani, in 2016 IEEE Int. Conf. Eng. Technol. (2016), pp. 112–116.
5
J. Han, M. Kamber, and J. Pei, Data Mining: Concepts and Techniques, Third (The Morgan Kaufmann, 2011).
6
I.B.N.S. Hardaya, A. Dhini, and I. Surjandari, in 2017 3rd Int. Conf. Sci. Inf. Technol. (2017), pp. 144–149.
7
K. Moolthaisong and W. Songpan, 2020 Int. Conf. Data Sci. Artif. Intell. Bus. Anal. DATABIA 2020 - Proc. 89
(2020).
8
A.M. Rahat, A. Kahir, and A.K.M. Masum, in 2019 8th Int. Conf. Syst. Model. Adv. Res. Trends (2019), pp. 266–
270.
9
G. Singh, B. Kumar, L. Gaur, and A. Tyagi, in 2019 Int. Conf. Autom. Comput. Technol. Manag. (2019), pp. 593–
596.
10
J.J. García Adeva, J.M. Pikatza Atxa, M. Ubeda Carrillo, and E. Ansuategi Zengotitabengoa, Expert Syst. Appl.
41, 1498 (2014).
11
C.E. Noviyanti and Y. Ruldeviyani, in 2020 Int. Work. Big Data Inf. Secur. (2020), pp. 29–34.
12
Noviantho, S.M. Isa, and L. Ashianti, in 2017 1st Int. Conf. Informatics Comput. Sci. (2017), pp. 241–246.
13
D. Buzic and J. Dobsa, 2018 41st Int. Conv. Inf. Commun. Technol. Electron. Microelectron. MIPRO 2018 -
Proc. 1011 (2018).
090003-6
Predict the Victory of the Freefire Esport Master
League Tournament Using Naïve Bayes
Heru Pusponugroho,1, a) Arief Setyanto,1, b) and Anggit Dwi Hartanto1, c)
Author Affiliations
1)
Magister Information Technic Universitas Amikom Yogyakarta, Yogyakarta, Indonesia
Author Emails
a)
Corresponding author: heru.1227@students.amikom.acid
b)
arief_s@amikom.ac.id.
c)
anggit@amikom.ac.id
Abstract. Esports is one of the biggest sports games in the world. Games are expanding in Indonesia as many events are held
nationally and internationally on youtube channel. The game that is now playing is Mobile Legend, PUBG, League Of Legend and
Garena Freefire . Freefire was held in 2020 throught tournament system that was participated by 18 clubs Esports divided into 3
groups of 6 team. The Esports tournament will go throught the regular season and playoff group. The problem that researchers is
to determine the prediction of victory throught several research method on Freefire Esports data sources. Therefore, the researcher
will use the method for Freefire Esport data namely Naïve Bayes. Researchers use this method becauseNaïve Bayes can determine
the data to find the highest and lowest team scores. The value that the researcher takes is based on the standings of the points that
pass are taken from the results of the calculation of this Naïve Bayes Method. The purpose of this paper is to obtain accurate and
normal datasets.
INTRODUCTION
Technological skill indeed brings many repercussions from positive and negative impact.if we can weigh the positives
and negatives, there may be many positive things, but for everyone there are also those who think that there are also
those who think that there are more negatives than positives depending on the assessment of each individual. The
communication behavior of Freefire online game users between teams gives each other the habit of giving each other
newswhen they want to play, thus making communication more intense and increasing the number of Freefire online
game users (Arya). In the era of increasingly advanced technology, the sports industry that is followed is Esports.
Freefire is one of the most popular sports games in southeast asia, including Indonesia. This game uses a supporting
smartphone in playing games. The most popular and frequently contested games both at national and international
levels . According to (zhang) Naïve Bayes is an algorithm that is often used is categorizing text. The basic idea is to
combine probabilities from document categories. Several previous studies that examined the prediction of the number
of wins in Esport game tournaments, among others by (santoso) using the Naïve Bayes method to predict the
percentage of wins based on the composition of the team they have, (Romadhon) predict percentage of general election
victory and percentage of athlete performance related to competitive advantage using SVM method where this method
is able to produce accuracy above 90% compared to other methods. In his journal, (Rian Ardiyanto) obtained the
results of the research, namely the Naïve Bayes algorithm obtained an accuracy value of 70.32%, precision 64.41%,
and a recall result of 90.36%. Algorithm support is optimized with the Smote method which has an accuracy value of
66.92%, a precision of 61.31%, and a recall value of 92.15%. So the method used is that the Naïve Bayes method is
superior to the methods of other researchers. Journal of (Windu Gata) related research on “Algorithm Implementations
Naïve Bayes, Random Forest. C4.5 on Online Gaming for Learning Achievement Predictions ”the accuracy value
obtained is 64.51% and the AUC value is 0.749 so that it gets a good accuracy value. Meanwhile, C4.5 got 67.58%
and the AUC value was 0.691. So it can be seen that the Naïve Bayes method has a higher value than the C4.5 method.
(Eduardus) “Dota 2 Esport Victory Prediction Based On Match Data “ the results of his research are 55,80% accuracy
with 32,41% precision and 46,70% recall. Overall, the accuracy of Naïve Bayes data for 15 Dota 2 match game data
is quite good, which is 80%. (Nurani) Research with the title "Comparison of the Performance of the Naïve Bayes
Algorithm and C4.5 for Food Price Classification" Results of data obtained by method researchers. C4.5 has an
accuracy rate of 5% better than the naïve Bayes method which is 65%. However, seen from the value of precision,
recall and accuracy for each method, only the difference is not much different. The aim of this research is that this
writing can find out the ranking data through booyah ranking and the most kills on each of these teams and to take a
sample of the Esports FreeFire club ranking in the playoffs using the Naïve Bayes method. This research method is
expected to accelerate the data processing in the standings. (Kusrini) research “ Analysis Of Prediction Of Students
Discipline level Using The NaïveConference
The 2nd International Bayes ofClassifier
Science and Algorithm Case Study
Information Technology in Smart :Administration
SMKNegeri 1Pacitan”
(ICSINTESA 2021) the result of this
study are indiscipline predictionAIPusing the Naïve
Conf. Proc. Bayes classifier
2658, 090004-1–090004-5; algorithm produce the best accuracy of 79,01%
https://doi.org/10.1063/5.0106992
090004-1
Second-level heading: Research Diagram
In the research diagram, there are several stages of the research in order to facilitate the process of searching for
Esport Freefire data using Naïve Bayes. The following are the research stages in figure below.
Stages of the Research diagram
Third-level heading: Example data Freefire
Based on picture above, there are .the stage of the research flow series that the researcher has made. Datacollection is
the initial stage for collecting through research papers that discuss the methods used to be discussed in this paper, as
well as applying methods that make it easier for researchers to calculate the amount of Esports FreeFire standings
data. The next stage is data collection, the researchers took this data from social media in the form of Instagram and
YouTube broadcasts to make it easier to find Esports FreeFire data as in Figure 2.1 below.
GRUP A GRUP B
Team M P Team M P
ONIC OLYMPUS 2 34 AURA ESPORTS 1 20
DRANIX VENDETTA 2 30 ISLAND OF GODS 1 17
BOSS TITAN 2 27 AEROWOLF PRO TEAM 1 14

FIRST RAIDERS
BRAVO 2 18 LOUVRE KING 1 10
BOOM ESPORTS 2 18 ONIC ARES 1 7
SAUDARA ESPORTS 2 17 DG ETERNAL 1 4
GRUP C
Team M P
EVOS ESPORTS 1 20
BIGETRON BIT 1 17
REDBULL REBELLION 1 14
BOSS KNIGHTMARE 1 10
RRQ HADES 1 7
DG ESPORTS 1 4
Freefire group stage standings data in 2020
090004-2
The data that is like in the picture above is the format of the Esport FreeFire Master League tournament andresearchers
will take it through the data in the FreeFire Master League tournament in the 2021 season where the data will be taken
by researchers via Instagram and YouTube broadcasts to rank all standings and put them together in the ranking playoffs
from groups A to C. Researchers will find it easy to rank the playoffs through the team that gets the mostbounces and
kills.
The next selection of ranking data is to take the highest value data that researchers take based on the results of the value
table data to be selected. Next, there is data processing, namely data that is done through the Naïve Bayes method so that
the results of the data get data accuracy that has reached a higher accuracy value so that the data process runs smoothly.
Furthermore, namely the Victory Data Accuracy Test, this data accuracy test tests the expected results of the data model.
The results from the table will be calculated from all the acquisition of the most kills and the total points in the standings.
Finally, the analysis is that the data that has been processed and calculated through the Naïve Bayes method will get the
highest value results and conclude from the data that has been entered.
MATH AND EQUATIONS
Method problems with the Naïve Bayes model can be described as follows. If given a data set S = {(‫ݔ‬,)} where each
sample is ‫ܰ ݐ‬ൌͳ(‫ݔ‬,) consists of data variables represented by m features (,, ...,), and the target variables {,, ...,} are
categorical data. The Naïve Bayes model is a model for predicting new data categories X = (,, ...,) using the conditional
independence probability which is formulated with the following equation: ‫א ݐ݉ݔʹݔͳݔݔݐ‬
ܿͳܿʹܿ݇‫݉ݔʹݔͳݔ‬
ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫ ݉ݔ‬ሃܿ݇ሻሺܿ݇ሻ
P (│,,…,) =ܿ݇ͳʹ݉
‫ݔ ݔ ݔ‬ (1)
ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫݉ݔ‬ሻ
Where: P factor (│,,…,) or posterior is the probability of the categoryܿ݇‫݇ܥ݉ݔʹݔͳݔ‬If the data variables are,,…,; or the
likelihood of the data attribute, ..., if it is known as the P () or prior data category, is the probability of the category
before processing the data and) is called the marginal probability.
‫݉ݔʹݔͳݔ‬ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫݉ݔ‬ሃܿ݇ሻ‫݇ܥ݇ܿ݇ܥ݉ݔ ͳݔ‬ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫݉ݔ‬ሻ (2)
Multiline equations
is considered constant, so the equation can be simplified to become:

ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫ ݉ݔ‬P (│,,…,) α P () ܿ݇‫݉ݔʹݔͳݔ‬ሺ‫ͳݔ‬ǡ‫ʹݔ‬ǡǥǡ‫݉ݔ‬ሃܿ݇ሻܿ݇ (3)
The Naïve Bayes model has two assumptions, namely:
1) In each data class (category), each data input attribute is statistically independent of each other.‫ݔ‬
ʹሻ P (│) of each feature j and data I have a certain distribution. Adjusted to the data attributes, the distributions
commonly used are: Gaussian (Normal), Multinomial, or Bernoulli distributions. For example: The Bernoulli
distribution is used if the feature is binary data, the multinomial distribution is used if the data is multi-category and
the Gaussian distribution is used if the data is continuous. According to the distribution used, there are three variances
of the Naïve Bayes classification model, namely the Naïve Bayes model for binary features, discrete features, and
continuous features.‫ݔ݇ܥݔ‬
090004-3
FIGURES AND TABLES
In this chapter, the researcher will try to take the standings that the researcher has taken on the YouTube Freefire account.
The account that the researcher has taken will be tested for the accuracy of the predictive victory data with the number
of 12 teams in the standings and calculate all data and predict who gets the highest, moderate, and decreasing score based
on the method the researcher made. The dataset that the researchers took was on the standings day 6 of the Group B and
C FreeFire Tournaments.The following is the FreeFire Tournament standings table as follows:
PLACE KILL TOTAL

RANK TEAMS PTS PTS PTS
1 RRQ HADES 51 35 86
2 ISLAND OF GODS 50 26 76
3 AURA ESPORTS 38 26 64
THEPRIME
4 ESPORTS 29 27 56
GGWP.JD
5 ESPORTS 22 29 51
6 FIRST RAIDERS 32 15 47
7 DG ESPORTS 29 18 47
8 EVOS ESPORTS 24 19 43
DRANIX
9 ESPORTS 19 21 40
10 KINGS ESPORTS 21 7 28
REDBULL
11 REBELLION 16 10 26
12 ROSUGO VEDA 11 5 16
Table of FreeFire Esport Standings Day 6 Tournament
After the researchers took the 12 teams, they would be divided into 3 parts, namely high, medium, and low scores. The
high value the researcher took from rank 1 to 5 because it got the highest total points, while the medium value was
taken at positions 6 to 9 and finally the lowest score was taken at positions 10, 11, and 12 because it got the lowest score.
From the results of the standings, it can be described using Naïve Bayes and the results are as follows:
‫ͳܥ‬: High
‫ʹܥ‬: Moderate
‫͵ܥ‬: Low
P () = 5/12 = 0.41 ‫ͳܥ‬
P () = 4/12 = 0.33 ‫ʹܥ‬
P () = 3/12 = 0.25 ‫͵ܥ‬
1. P (X │), i = "High", "Medium", "Low"‫ܥ‬
* High Class
a) P (Rank = "RRQ Hades" │ (5)) = 5/86 = 0.58
P (Rank = "Island Of Gods" │ (5)) = 5/76 = 0.65
P (Rank = “Aura Esports” │ (5)) = 5/64 = 0.78
P (Rank = “The Prime Esports” │ (5)) = 5/56 = 0.89
P (Rank = “GGWP ID Esports” │ (5)) = 5/51 = 0.98
b) P (Place = “51 - 50” │ (2)) = 2/5 = 0.4
c) P (Kill = “35, 27, 29” │ (3)) = 3/5 = 0.6
* Medium Class
a) P (Rank = “First Raiders” │ (4)) = 4/47 = 0.85
P (Rank = “DG Esports” │ (4)) = 4/47 = 0.85
P (Rank = “Evos Esports” │ (4)) = 4/43 = 0.93
P (Rank = “Dranix Esports” │ (4)) = 4/40 = 0.1
b)P (Place = “32, 29” │ (2)) = 2/4 = 0.5
c) P (Kill = “18, 21” │ (2)) = 2/4 = 0.5
090004-4
* Low Class
a) P (Rank = “Kings Esports” │ (3)) = 3/28 = 0.10
P (Rank = "RedBull Rebellion" │ (3)) = 3/26 = 0.11
P (Rank = "Rosugo Veda" │ (3)) = 3/16 = 0.18
b)P (Place = “11” │ (1)) = 1/4 = 0.25
c) P (Kill = “5, 7” │ (2)) = 2/4 = 0.5
2. The valued attribute:
a. Class: Team = "High"
P (X │) = 0.58 x 0.4 x 0.6 = 0.1392 "RRQ Hades"‫ͳܥ‬
P (X │) = 0.65 x 0.4 x 0.6 = 0.156 "Island Of Gods"‫ͳܥ‬
P (X │‫ = )ͳܥ‬0.78 x 0.4 x 0.6 = 0.1872 "Aura Esports"
P (X │‫ = )ͳܥ‬0.89 x 0.4 x 0.6 = 0.2136 "The Prime Esports"
P (X │‫ = )ͳܥ‬0.98 x 0.4 x 0.6 = 0.2352 "GGWP ID Esports"
b. Class: Team = "Medium"
P (X │) = 0.85 x 0.5 x 0.5 = 0.1875 "First Raiders"‫ʹܥ‬
P (X │) = 0.85 x 0.5 x 0.5 = 0.1875 "DG Esports"‫ʹܥ‬
P (X │‫ = )ʹܥ‬0.95 x 0.5 x 0.5 = 0.2375 "Evos Esports"
P (X │‫ = )ʹܥ‬0.1 x 0.5 x 0.5 = 0.25 "Dranix Esports"
a. Class: Team = "Low"
P (X │) = 0.10 x 0.25 x 0.5 = 0.125 "Kings Esports"‫͵ܥ‬
P (X │) = 0.11 x 0.25 x 0.5 = 0.1375 "Redbull Rebellion"‫͵ܥ‬
P (X │‫ = )͵ܥ‬0.18 x 0.25 x 0.5 = 0.225 "Rosugo Veda"
b. Based on the Place and Attribute Probability Value
P (X │P () = 0.4 x 0.41 = 0.164‫ͳܥ‬
P (X │P () = 0.5 x 0.33 = 0.165‫ʹܥ‬
P (X │P () = 0.25 x 0.25 = 0.625‫͵ܥ‬

CONCLUSION
Based on the result of data calculations using the Naïve Bayes method, it can be concluded that GGWP ID ESports gets
the high class with a value of 0.2352, while the medium class is obtained by Evos Esports with a value of 0.2375, and
the low score the researcher takes the few points, Kings Esports gets value 0.125. Researchers take Esport data based
on valuable attributes. Researchers also get the place and attribute probability value where the probability is obtained
with a value of 0.625 and the others get low results. This proves that the Naïve Bayes method is very easy in searching
for Esports FreeFire standings data even though the data is not optimal and is obtained by the GGWP ID ESports club.
ACKNOWLEDGMENTS
Researchers want to thank their parents and friends who helped with the journal arrangement. Despite many flaws in
research. Hopefully, what research have created can be used by others.
.
REFERENCES
1. Arya Barata, Achmad Rosihan. 2020 “Analysis Of online Freefire Game Communication Behavior In Students Of Vocational School 1
Martapura, East oku District” 2-3 http : //jourrnal.unbara.ac.id/index.php/JM/index.
2. zhang, w & Gao, f. 2011 “An Inprovement To Naïve Bayes For Text Classification. Procedia Engineering “ 1-2 Procedia Engineering
15 (2011) 2160-2164 .
3. Santoso, I., Gata, w., & Paryanti, AB .2019 “Use Of Feature Selection In The Support Vector Machine Algorithm For General Election
Commision Sentiment Analysis” Journal Of RESTI (System Engineering And Information Technology), 3 (3) , 364-370.
4. Ednawati Rainarli, Afif Romadhan.2017” Comparison Of Simple Logistic Classifier With Support Vector Machine In Predicting Athletes’s
Winning” Journal Of Information systems Engineering And Business Intelligence Vol 3., No 2.
5. Eduardus Hadika Atmaja, 2020, “Prediction Of Dota 2 Esport Victory Based On Competition Data”Informatics Study Program, Faculty Of
Science And Technology, Sanata Dharma University, Sleman, Yogyakarta.
6. Niken Puji astuti, Kusrini, M. Rudyanto Arief. 2015 “Analysis Of Prediction Of Undipline Students Using Naïve Bayes Classifier Algorithm
(Case study : SMKNegeri 1 Pacitan)” National Seminar On Information And Multimedia Technology, ISSN: 2302-3805 .
7. Lila Setyani, Mochhamad Wahidin., Et. al.2020 “ Analysis Of Timely Student Graduation Prediction Using Naïve Bayes Data Mining
Method: Systematic Review” Factor Exacta 13 (1) : 35-43, 2020, DOI : 10.30998 / Factorexacta.v13i1.5548.
8. Nurani., Afif, 2020 “ Performance Comparison Of Naïve Bayes Algorithm And C4.5 For Food Price Classification” 1-2 @ 2016 Protek :
Electrical Engineering Scientific Journal Quantities, ANSI Standard Y10.5-1968.
9. Yaya Heryadi, Teguh Waryono. 2020 “machine Learning (Konsep & Implementasi)” Gava Media, Yogyakarta.
090004-5
Analysis of the Application of Data Mining Clustering
Model in the Selection of the Best Study Program Based on
Student Interests at the University of Aceh Province
Yuli Asbar 1,a) and Sapna Biby1,b) , Janner Simarmata2,c) , Abror3,d) , Angga
Pratama4,e) , Mutammimul Ula4,f)
Author Affiliations
1
Department of Management, Faculty of Economics and Business, Universitas Malikussaleh, Aceh, Indonesia
2
Medan State University, Medan, Indonesia
3
Faculty of Economics, Universitas Negeri Padang, Padang, Indonesia
4
Department of Information System, Universitas Malikussaleh, Aceh, Indonesia
Author Emails
a)
Corresponding author: yuliasbar@unimal.ac.id
b)
sapnabiby@unimal.ac.id
c)
jannersimarmata@unimed.ac.id
d)
abror094@fe.unp.ac.id
e)
anggapratama@unimal.ac.id
f)
mutammimul@unimal.ac.id
Abstract. Limited student capacity at State Universities (PTN) in Aceh provides opportunities for Private
Universities (PTS) in Aceh to hold lectures at universities. It provides opportunities for private universities,
institutes, colleges, academies, and other polytechnics in student admissions. Marketing strategies in doing
promotions are needed to be able to attract students to enroll in the college. Instead, the promotion strategy must be
on target effectively and efficiently. Therefore, a model is needed in looking at the grouping of students who are
most widely accepted in the study program and the origin of the school. This is to facilitate the campus in carrying
out promotional strategies to schools that contribute the most students to private universities in Aceh such as al-
Muslim and uniki universities. The existence of this problem requires a system in looking at the grouping patterns
that are most widely accepted to private universities based on schools and study program interests. Furthermore, a
model is needed to make the student's interest strategy in looking at the study program in demand so that private
universities can prepare strategies in conducting promotions going forward and the funds allocated for promotion
can be used appropriately. Data taken at the time of the study can be seen from each school that contributed as much
students to historical data starting from 2015, 2016, 2017, 2018, 2019 2020, which then the data was taken from
each college. The results of this study can also look at the grouping of schools that contribute a lot of students to
each school and the pattern of grouping students most widely accepted by school and historical in previous years
using the optimal pattern of number of clusters with the average silhoutewidth method in looking at the distribution
of students who have been in college The results of the k-means research analysis of the application of the Data
Mining Clustering Model in the Selection of the Best Study Programs Based on Student Interests at the University,
the results of cluster 1: the minimum average in the number of students spending per year is 7 people, max 12
people, cluster 2: per year is 2 people, max 5 people, cluster 3: 26 people per year, max 31 people, cluster 4:
minimum average of 15 people, max 20 people.
090005-1
INTRODUCTION
The choice of majors at a university is an important decision in determining future education, this is how
the campus sees future opportunities 1. One of the methods in viewing majors by students chooses to use the
clustering technique. one of the main placements of students is very influential on students going forward and can
direct students better in improving achievement in their academic fields. The model used based on the student's
choice of majors is K-Nearest Neighbor (K-NN) and Naïve Bayes Classifier (NBC) which is then measured by
Matrix Accuracy 2. There is a fundamental problem in seeing student admissions from various schools.
Furthermore, the large number of data distributions from various schools opens up opportunities for student
admissions. Therefore, it is important to apply data mining techniques in looking at the management of the
number of students and the origin of students in an education, especially during the computer-based learning
period of intelligent systems (M. A. Nurrohmat,2015)3.
Furthermore, data mining can play an effective role in conducting real-time data analysis with very large
data volumes and being able to predict the future by looking at some of the information and existing data patterns
4,5
. In the selection of the suitability of majors there are several problems such as a lot of data that takes a long
time to search for data and energy and time is wasted. Required accuracy of data processing based on academic
potential is very important at the time of recapitulation for and requires extra precision. There are some students
and parents who ignore the accuracy of data processing based on academic potential, another problem is
determining the majors according to the wishes of students without paying attention to academic values, this
causes students to experience problems. Then many variables in determining the majors will affect the chosen
interests such as Mathematics, Natural Sciences, Social Sciences, and English2.
Data mining in determining the type of voice that there is a classification modeling model which then the
classification model that plays a role in the value grouping process, this model is also made in the classification
for promotional strategies in each school where there are many students to the university.6 Grouping majors at
each university using the k-means algorithm that has been tested in viewing clusters and grouping on market
patterns, pattern recognition, information seeking. The k-means algorithm has high performance in grouping the
previous data and can be implemented quickly at the time of grouping. 7,8.
LITERATURE REVIEW
Data Mining dan Clustering
The important thing now is that the data and the appropriate grouping is the one that has more value.
With the increase in magnitude it is interesting to get a pattern in this data. Therefore, the importance of how to
process and use data effectively is increasing. New techniques to help analyze, understand the large amount of
data in the database. Another term for data mining is the process of discovering significant things of knowledge,
such as patterns, associations, changes, anomalies, and significant structures from large amounts of data stored in
databases, data warehouses, or other information storage areas 9. One method that is often used is the Clustering
model where clustering is a technique used to partition data (objects) into separate and homogeneous groups
(clusters) of objects, so that objects in a cluster resemble objects in other clusters. Then there is a grouping of
algorithms from various fields including: machine learning, neural networks and statistics. 10,11. Priority in
determining majors becomes a strategy in selecting the best campus. This students can see from the grouping and
ranking of universities. Therefore, there is a clustering model model to clarify the presentation of the model in the
favorite major strategy that uses cash in a country 12. Another problem that is often faced is that there is no
historical database for students who have been accepted and there has not been a pattern or way to carry out a
suitable promotion in promoting new students 13.
New student admissions each year a lot of data is generated consisting of profiles of prospective students
which causes a lot of data accumulation in the system database. Another problem for the campus so far is that it is
difficult to identify students and schools from which many are accepted on the campus. because of the large
number of new student data the same data. With the clustering model using k-means clustering to identify
prospective students who have the same similarity character to make clusters in data mining 14. Interests and
talents greatly influence decisions in school selection. Therefore the promotion strategy of each school can be
seen by the talents and interests of these students. Cluster analysis with data mining has the same characteristics
resulting in grouping results based on the school's area of origin with the highest average GPA 15,16. The problem
that is often faced is that there is no historical database for students who have been accepted and there has not
been found a pattern or way to carry out a suitable promotion in promoting new students 13.
090005-2
The existence of designing and evaluating the implementation of information security governance (ISG)
in environments such as universities and banks to manage Strengths and weaknesses, information is seen from
document review. Furthermore, this strategy is very important to maintain student data in the historical database
in the previous year. So that no threat to the database is applied 17. The process of accepting new students has a
lot of data stored and the data causes a large buildup. Data mining model k-means clustering to identify
prospective students who have the same similarity character. The campus has been difficult to identify these
students. In the admission of new students there are the same amount of data. Therefore, the k-means clustering
model of data mining is used. The largest value cluster is accepted in the science and social studies majors, on the
other hand, if it is small, it is rejected with the smallest centroid value 14,18. Interests and talents greatly influence
student decisions in school selection. Therefore the promotion strategy of each school can be seen by the talents
and interests of these students. Cluster analysis with data mining has the same characteristics resulting in
grouping results based on the school's area of origin with the highest average GPA 15,16.
Promotional Marketing Strategy
Marketing promotion strategies at each school can direct students to focus more on developing their
abilities and interests in majors. With this research, it is expected to maximize individual potential, talent or
talent, so as to maximize academic value and the existence of large amounts of data based on school history.
Hence the K-Means Algorithm. used in determining majors from various criteria used in grouping student data at
each university, this makes it easier to carry out future promotions 19. Determination of the type of criteria using
the Simple Multi Attribute Rating Technique Exploiting Rank (SMARTER) method and the Simple Additive
Weighting (SAW) method in determining soil type 20. In the promotion strategy with the use of information
technology currently provides a lot of convenience, speed up, and streamline work. The existence of
technological developments with RFID makes it easier for lecturers and students to monitor progress and the
existence of a database will store all activities including promotional strategies that have been stored in history 21.
Subsequent research analyzes the factors that influence purchasing decisions of online pre-order systems. The
results of the research are that there are variables that influence each other and then marketing strategies can be
applied to students in seeing the convenience, trust, and reputation of the campus have a significant effect on
student decisions 22.
Subsequent research analyzes the factors that influence purchasing decisions of online pre-order systems.
The results of the research are that there are variables that influence each other and then marketing strategies can
be applied to students in seeing the convenience, trust, and reputation of the campus have a significant effect on
student decisions 23. Implementation of the K-means Clustering Algorithm to Determine Marketing Strategy at
the University students come from various regions in Indonesia. marketing strategy in conducting promotions to
find prospective students so that promotions are carried out more effectively and efficiently 24. One of the
promotional strategies, especially for universities, is the existence of a website/site to view information so that it
can reach new students. Use through the use of social media, enabling integrated marketing activities with lower
effort and costs than ever before. Promotion from an external perspective through referrals and companies can
use several kinds of promotional mix (reference group), family (family) and marketing activities (marketing
activities), publicity (Public relations), personal selling (Personal selling), and direct marketing (Direct
marketing), (Kotler, 2002:644). 25–27.
RESEARCH METHODOLOGY
1. The First Stage of the Business Understanding Phase
Business understanding is done by studying the object of research, namely the University from which data will be
taken based on historical data. The purpose of grouping students based on the listed schools of origin.
2. Data Comprehension Phase
The data used for the analysis process of the K-Means method is student history data that will be grouped.
Historical data will be clustered at the origin of the school whose data after being obtained, the next process is to
understand the data.
3. Data Processing Stage (Data Preparation Stage)
The data preparation stage is to select the attributes that will be used for modeling with the K-means clustering
process model taken from student data. Furthermore, the attributes are stored in the database which will then be
processed at the modeling stage
4. Modeling Phase
For this stage, the clustering method process will be carried out with a clustering algorithm with K-means
modeling tools using a desktop application
090005-3
The following Application of Data Mining Clustering Model in The Selection of The Best Study
Program Based on Student Interests at the University as follows :
END
Problem identification and business

understanding phase
Analyzing the Problem
Data Comprehension Phase
Modeling and Analyzing Phase of the K-

Means Clustering Algorithm
K-Means Clustering
Results
Calculate Silhoutte Coefficient
Silhoutte Coefficient
Results
END
FIGURE 1. Research Methodology

In this study, there were 107 schools that enrolled students in all of the Lhokseumawe and North Aceh
districts. Data on the average achievement of the program for the year each school 2010-2021.
TABLE 1. Initial Data Model
NPSN SEKOLAH 2010 2011 2012 2013 2014 2017 2018 2019 2020 2021
10105619 Sman 1 Lhokseumawe 20 22 26 … … 29 27 28 30 28
10105604 Sman Modal Bangsa Arun 32 19 33 … … 26 22 26 30 34
11111112 Sma Swasta Sukma Bangsa 13 15 16 … … 13 15 17 20 18
11111113 Sma Swasta Yapena 20 19 13 … … 15 20 14 17 19
11111114 Smkn 1 Lhokseumawe 34 35 30 … … 29 27 30 34 29
10101174 Sman 1 Lhoksukon 27 25 29 … … 27 25 23 28 31
10101175 Sman 1 Dewantara 20 32 34 … … 25 27 23 28 31
10101176 Sman 1 Cot Girek 14 14 13 … … 14 18 15 16 20
10101181 Sman 2 Lhoksukon 20 17 17 … … 19 18 17 13 18
10101183 Sman 1 Nisam 15 8 13 … … 12 10 8 9 8
… … … … … … … … … … …
10101309 Sman 1 Matangkuli 2 3 5 … … 5 3 2 2 4
10101311 Sman 1 Kutamakmur 9 10 8 … … 9 8 11 9 8
10101312 Sman 2 Dewantara 14 15 14 … … 15 16 17 19 20
10106274 Sman 2 Nisam 8 9 11 … … 10 9 8 8 9
10106277 Sman 2 Seunuddon 9 8 8 … … 12 9 8 8 9
10108100 Sman 1 Muara Batu 6 8 6 … … 6 8 8 10 8
11111141 Man Lhoksukon 14 16 17 … … 16 14 13 13 16
11111142 Man Matangkuli 8 10 12 … … 9 12 11 9 12
090005-4
NPSN SEKOLAH 2010 2011 2012 2013 2014 2017 2018 2019 2020 2021
11111161 Smkn 1 Muara Batu 9 8 9 … … 7 8 8 10 8
11111162 Smk N 1 Dewantara 10 10 12 … … 9 10 9 8 11
11111163 Smk N 1 Lhoksukon 9 12 11 … … 12 16 12 14 16
2. Data Hasil Cluster K-Means

In this study, there were 107 schools that enrolled students in all of the Lhokseumawe and North Aceh
districts. Data on the average achievement of the program for the year each school 2010-2021 adalah sebagai
berikut:
FIGURE 2. K-Means Cluster Result Data

3. Data Cluster Plot
Data Cluster Plot Analysis of The Application of Data Mining Clustering Model in The Selection of The Best
Study Program Based on Student Interests at the University of Aceh Province
FIGURE 3. Cluster Plot Data

3. Cluster optimal number Average Silhouette width
Data Cluster Plot Analysis of The Application of Data Mining Clustering Model in The Selection of The Best
Study Program Based on Student Interests at the University of Aceh Province
FIGURE. 4. Cluster Optimal Number Average Silhouette Width
090005-5
CONCLUSION
Based on the results of research on the Analysis of the Application of Data Mining Clustering Models in
the Selection of the Best Study Programs Based on Student Interests, it can be concluded that the following
conclusions can be drawn:
1. With this system, the campus knows the most schools that have students based on the results of the k-
means clustering, then there are schools that are medium and few. This makes it easier for the team to carry
out promotions in each school and the efficiency of time is more efficient and effective
2. The results of K-means clustering for the distance between the centers of clusters C1 and C2 with a
distance of 5.918444803, then for clusters C1 and C3 with a distance of 5.42117556, then for C2 and C3
with a distance of 1.343755104, the BCV value is 12.68337547. the results seen in each cluster are the
current cluster ratio 0.300791513 and the previous ratio 0.300791513 then the current ratio is not > from
the previous ratio, then the iteration is stopped. with the results of clustering SMA 1,2,3, SMA DEwantara
Lhokseumawe Cluster 1 with the most students, for the medium cluster Cluster 2 and Cluster 3 Next which
is following the smaller cluster below it.
ACKNOWLEDGMENTS
Thanks to the research team and the school office, vice chancellor 1 and academic assistants in each faculty and the
education office for helping this research run smoothly and successfully
.
REFERENCES
1. Porter, S. R., & Umbach, P. D. College major choice: An analysis of person–environment fit. Research in
higher education. (2006).
2. Hastarimasuci R, Papilo P, N. A. Variable Selection to Determine Majors of Student using K-Nearest
Neighbor and Naïve Bayes Classifier Algorithm. J. Phys. Conf. Ser. 1363, (2019).
3. M. A. Nurrohmat. Aplikasi Pemrediksi Masa Studi dan Predikat Kelulusan Mahasiswa Informatika
Universitas Muhammadiyah Surakarta Menggunakan Metode Naive Bayes. Khazanah Inform. J. Ilmu
Komput. dan Inf. 1, 29–34 (2015).
4. J. Archenaa and E. A. M. Anita. A survey of big data analytics in healthcare and government. Procedia
Comput. Sci 50, 408–413 (2015).
5. Ramageri, M. B. & M. Data mining techniques and applications. (2010).
6. Rosdiana, R., Ula, M., & Aidilof, H. A. K. Implementasi Pemodelan Citra Model Svm (Support Vector
Machine) Dalam Penentuan Pengklasifikasian Jenis Suara Kontes Burung. J. Inform. Kaputama 317–324
(2021).
7. M., C. Y. K-means: A new generalized K-means clustering algorithm. Pattern Recognit. Lett. 24, 2883–
2893 (2003).
8. Kuo R. J. Ho L. M., & H. C. M. Integration of self-organizing feature map and K-means algorithm for
market segmentation. Comput. Oper. Res. 29(11), 1475–1493 (2002).
9. Liao S. H., Chen C. M., & Wu, C. H. Mining customer knowledge for product line and brand extension in
retailing. Expert Syst. Appl. 35(3), 1763–1776 (2008).
10. Boutsinas B., & G. T. On distributing the clustering process. Pattern Recognit. Lett. 23(8), 999–1008
(2002).
11. Oscar Corcho, Mariano Fernández-López, A. G.-P. Methodologies, tools and languages for building
ontologies. Where is their meeting point? Data Knowl. Eng. 46, 41–64 (2003).
12. Rad A., Nader B., & S. M. Clustering and ranking university majors using data mining and AHP algorithms:
A case study in Iran. Expert Syst. Appl. 38(1), 755–763 (2011).
13. Dana, Raditya Danar, Cep Lukman Rohmat, A. R. R. Strategi Marketing Penerimaan Mahasiswa Baru
Menggunakan Machine Learning Dengan Teknik Clustering. J. Pengemb. IT 04(2), 201–204 (2018).
14. Setiawan, R. Penerapan Data Mining Menggunakan Algoritma K-Means Clustering Untuk Menentukan
Strategi Promosi Mahasiswa Baru Baru (Studi Kasus: Politeknik Lp3i Jakarta). J. Lentera Ict 3(1), 76–92
090005-6
(2017).
15. Purwanto A., & I. & E., R. menentukan strategi promosi politeknik tmkm berdasarkan analisa pengaruh
minat siswa. Pros. Semnastek (2017).
16. T., S. Klasifikasi data mahasiswa menggunakan metode k-means untuk menunjang pemilihan strategi
pemasaran. JIKO (Jurnal Inform. dan Komputer) 1(1), (2016).
17. Ula M., F. W. A method for evaluating information security governance (ISG) components in banking
environment. J. Phys. Conf. Ser. 812, (2017).
18. Purnamaningsih C., Saptono R., A. A. Pemanfaatan Metode K-Means Clustering dalam Penentuan
Penjurusan Siswa SMA. ITSMART. J. Teknol. dan Inf. 3(1), 27–33 (2014).
19. Y., I. Implementation Of Data Mining For Determining Majors Using K-Means Algorithm In Students Of
SMA Negeri 1 Pangkalan Kerinci. J. Appl. Eng. Technol. Sci. 1(1), 17–29 (2019).
20. Nurdin N., Fahrozi F., & U. M. Sistem Pendukung Keputusan Penentuan Jenis Tanah Yang Sesuai Untuk
Tanaman Pangan Menggunakan Metode Smarter Dan Saw. Inform. Pertan. 29(2), 83–94 (2020).
21. Ula M., Pratama A., Asbar Y., Fuadi W., Fajri R., H. & R. A New Model of The Student Attendance
Monitoring System Using RFID Technology. J. Phys. Conf. Ser. 1807, (2021).
22. Asbar Y., S. & V., W. Analysis of purchasing decisions using online pre-order system at Shopee. Oper.
Manag. Inf. Syst. Stud. 1(1), 30–40 (2021).
23. Aranda J., N. W. A. G. Penerapan Metode K-Means Cluster Analysis pada Sistem Pendukung Keputusan
Pemilihan Konsentrasi Untuk Mahasiswa International Class STMIK Amikom Yogyakarta. in Seminar
Nasional Teknologi Informasi dan Multimedia 2016.
24. O., O. J. Implementasi Algoritma K-Means Clustering Untuk Menentukan Strategi Marketing President
University. J. Ilm. Tek. Ind. 12(1), 10–20 (2013).
25. Kim Angella J, E. K. Do social media marketing activities enhance customer equity? An empirical study of
luxury fashion brand. J. Bus. Res. No.65, 1480−1486 (2012).
26. Hawkins N., Richards P., Granley H., S. & D. The impact of exposure to the thin-ideal media image on
women. Eating Disorders. J. Treat. Prev. 12(1), 35–50 (2004).
27. Amstrong, K. dan. Prinsip-Prinsip Pemasaran. (Erlangga, 2008).
090005-7
A Modification of Decomposition Approach for Solving Non-
linear Quadratic Differential Equation; Theory and
Application
Khalid Hammood AL-Jizani1, a) and Suhartati2, b)
Author Affiliations
1
Department of Mathematics, College of Science, Mustansiriyah University, Baghdad, Iraq
2
Department Information System, Computer Science, Mulia University, Balikpapan, Indonesia
Author Emails
a)
Corresponding author: khalid.math79@uomustansiriyah.edu.iq
b)
suhartati@universitasmulia.ac.id
Abstract. In this study, we introduce a modified Adomian decomposition (MADM) for solving non-linear quadratic form.
A solution for these types conducted via a classical Adomian decomposition approach (ADM) award perfect analytic
solution just on neighborhood to initially place. A main feature for the proposed technique, it give us the ability to increase
area of convergent approximation solution. A solution conducted via MADM award nice approximate results to wider and
bigger intervals, as well as previously initially place. Selected examples confirm the proposed technique is simply, more
an accurate and effectiveness.
PRELIMINARIES
A modify to Adomian decomposition approach (MADM) discussed to address a quadratic form:
uc(t ) b(t ) r (t )u (t ) o(t )u 2 (t ),

°
® 0dt dT (1)
°̄u (0) E ,
When b(s), r (s) and o( s) are continuous functions, in addition more important in different fields science. For
instance, as is famously, a single-dimensional type very close with a quadratic form. Motivation for the proposed
technique is to get a highly an accurate solution through expand more extended of the convergence region and it gives
the solution in infinite series. The most important advantages of this technique has a capability to provide a
continuously represent for the analytic solution and allow best details result with whole interval time. Other feature to
MADM it overhead an exist approaches to solve these problems, is a solution for equation (1) conducted via proposed
method is efficiently not just too small values x but to larger values.
LITERATURE REVIEW
An exact solutions difficult for obtaining, therefore many approximate analytical methods appeared for this
purpose 1, 2, 3, 4, 5, 6, 7.
At 1980, Adomian technique (ADM) is an example of approximate analytical method which successfully
implemented on different kinds including nonlinear or linear ordinary and partial differential equations. This technique
implemented on large types of equation form 8, 9. A mainly feature for ADM, it may to apply direct on different types,
100001-1
non-linear with non-homogeneous, with constants or with variable constant10, related to accurate with rapidly
convergent 10, 11. It is implemented for solving Riccati differential type 12, 13. Author implemented a multistep Adomian
technique to solve Riccati types then comparison result with classical ADM14, 15. Also, for 16, solved delay diơerential
equation problems using ADM, while in 13, have used ADM to solve nonlinear Sine Gordon type. A solution for the
proposed technique has a rapidly convergent to closed form in general. Some modify for this approach shown 17.
METHODOLOGY
Analysis of Adomian Decomposition Method (Adm) 8, 18, 7
For general equations, let us to take:

Lu Ru Nu g (2)
With prescribed conditions, where u is the unknown function, L is high derivation is supposed which easy invertible,
R is a linear diơerential operator of order less than L, N is a non-linear part, g is the source part. Implementing
inversely operator L1 for this equation (2), with given initial condition, one get:
u f L1 ( Ru) L1 ( Nu) (3)
݂ Is a term which arise by integration of source term g with use the initial or boundary condition. A classical Adomian
approach define a solution u ( x) as a series:
f
(4)
u (t ) ¦u
n 0
n (t )
Then decomposition a non-linear N as:

f
Nu ¦A n 0
n
(5)
An Is Adomian polynomial u0 , u1 , u2 ,... .
A traditional form of Adomian polynomial is:

1 ª dn § f ·º (6)
An «
n! « d O n
N¨
¨
©
¦O t i
i ¸»
¸»
¹¼ O
¬ i 0 0
Equation (6) can be calculate via Mathematic technique. Substituting (4) and (5) into (3) yields:
f § f · § f · (7)
¦u
n 0
n (t ) f L1 ¨ R
¨
© n 0
¦
un (t ) ¸ L1 ¨
¸
¹
¨
©n 0
An ¸
¸
¹
¦
Thus, we can identify:
u0 (t ) f (t )
(8)
un 1 (t ) L1 ( Run ) L1 ( An ), n t 0
Finally, we approximate the solution by the M th term:

f
M M (u ) ¦u
n 0
n
With (9)
lim MM (u ) u
M of
100001-2
Theorem 1 (Banach Theorem)
Suppose B Banach space with
D: B o B
Is non-linear mapping, with assume
D[u] D[w] d E u w , u , w B
For some constants E 1. then A has a unique fixed point. Also, the sequence
un1 D[un ],
With an arbitrary selection for u0 B, converge to fixed point of D. Based on Theorem 1, for the non-linear
mapping
D[u(t )] L1( Run ) L1( An ) (10)
A necessary condition for convergence of ADM is strictly contraction of D. In addition, equation converge to the fixed
point for D which is also a solution equation. In addition, equation (8) converge to the foxed point for D which is also
a solution equation (2).
Modification Adomian Decomposition Technique to Solve 2

A mainly obstacle for traditional ADM is that a solution which conducted May rapid converge but for a smaller
regions. For increasing a convergent area to results are conducted, one must modify ADM via introduce auxiliary
parameter J . where J as a parameter such that J z 0, which use to regulate a convergent regions. Related to
equation (8), it will become as:
un1 (t ) J [ L1 ( Run ) L1 ( An )], n t 0 (11)
In fact, a formula (10) give big freedom of select. Related to complex non-linear equations, we may choose a
smaller value of J (in general below one) to conduct a better analytic solution with wide regions. Additionally, we
may observe when J = 1, equation (11) becomes the standard Adomian formula (8).
Theorem 2
Suppose that u0 (t ) E with a sequence {un (t )} conducted by (11) converge to u (t ); then u (t ) be the solution of
Eq.1.
Proof. Taking limits to the both sides of Eq. 11, we get:
lim un 1 (t ) J lim [ L1 ( Run ) L1 ( An )], n t 0 (12)

mof mof
And thus
J lim [ L1 ( Run ) L1 ( An )] 0. (13)
m of
Since J z 0, it conclude that:
[ L1 ( Run ) L1 ( An )] 0. (14)
That’s mean the differential equation for the both sides respect to t yield:
u c(t ) b(t ) r (t )u (t ) o(t )u 2 (t ). (15)
Means u (t ) satisfy Eq.1 and u (0) E since un (0) E . Mean u (t ) a solution of equation (1) which completely
proof. Based on Banach theorem, easily to get the convergent condition to a sequence u n conducted via (11).
100001-3
Theorem 3
A define of non-linear mapping:

T [u(t )] J [ L1 ( Run ) L1 ( An )], n t 0.
A sufficient condition to convergent of {un (t )} conducted from (11) strictly contraction of non-linear mapping T.
Additionally, equation (11) converge to a fixed point of T which is also a solution to equation (1). Then based on (11),
selecting adequate J with u0 (t ), an analytic solution to (1) on whole [0, T ] can be obtained.
NUMERICAL EXPERIMENTS
Now, apply the MADM discussed to different types of quadratic differential equations. Examples shown a MADM
is very simply, more an accurate and effectiveness.
Example 1
Consider the quadratic DE
u c(t )
° 1 t 2 u 2 (t ), 0 d t d 4,
®
°̄u (0) 1,
Where a closed form is (16)
t 2
e
u (t ) t t
.
³e
z2
1 dz
0
According to (10) choosing J 0.8, m 15, it can obtain a best approximations on t 0, 4 Table 1, show all results and
more clearly from the figures. It is easily to see when one choose the smaller value of J, can get a wider for the
convergence region for the sequence (11).
TABLE 1. COMPARISON BETWEEN THE EXACT AND APPROXIMATE SOLUTION
t MADM Exact Solution Absolute Error

0.1 1.00031731031038 1.00031731031038 1.33227 ×10-15
0.2 1.00241982548045 1.00241982548044 3.9968×10-15
0.3 1.00779458808214 1.00779458808213 9.54792 ×10-15
0.4 1.01765087886187 1.01765087886255 6.82343×10-13
0.5 1.03295757449042 1.03295757602655 1.53613×10-9
0.6 1.05446646377857 1.05446680992334 3.46145×10-7
0.7 1.08270318465693 1.08272748135379 2.42967×10-5
0.8 1.11725224444244 1.11809254540645 8.40301×10-4
0.9 1.14260854260619 1.16072397273281 1.81154×10-2
1. 0.93136253864647 1.210559901466964 2.79236×10-1
100001-4
1.12
1.10
1.08
1.06
1.04
1.02
0.2 0.4 0.6 0.8
FIGURE 1. The Exact and Approximate Solutions of x15 (t )
Example 2
Consider the quadratic DE

° y (t ) y 2 (t ) 1,
® 0 d t d 6, (17)
°̄ y (0) 0,
Where the exact solution is given by:

e 2t 1
y (t ) .
e 2t 1
According to (10), choosing J 0.9, m 30, it can obtain a best approximations on t 0,6 Table 2, show all results
and more clearly from the figures. It is easily to see when one choose the smaller value of J, can get a wider for the
convergence region for the sequence (11).
TABLE 2. COMPARISON BETWEEN THE EXACT AND APPROXIMATE SOLUTION
t MADM Exact Solution Absolute Error

0.1 0.099667994624955 0.099667994624955 1.3778 ×10-17
0.2 0.197375320224904 0.197375320224905 2.77556×10-17
0.3 0.291312612451591 0.291312612451591 0
0.4 0.379948962255225 0.379948962255225 5.5112 ×10-17
0.5 0.46211715726001 0.46211715726001 5.5112×10-16
0.6 0.537049566998039 0.537049566998035 3.33067×10-15
0.7 0.604367777117171 0.604367777117164 6.88338×10-15
0.8 0.664036770248793 0.664036770267849 1.90556×10-11
0.9 0.716297865149588 0.716297870199024 5.04944×10-9
1. 0.761593613986952 0.761594155955765 5.41969×10-7
100001-5
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.2 0.4 0.6 0.8
FIGURE 2. The exact and approximate solutions x15 (t )
CONCLUSION
In this study, a MADM is discussed for solving quadratic form. Comparison with the exact solution is conducted,
result related to examples explain a proposed technique may award more an accurate solution with wider regions,
which a mainly feature of the proposed approach. Furthermore, a modification of ADM may overhead a restriction to
application area of standard ADM. In general, when J selected a smaller, get an accurate results and more extended
of the convergence region.
ACKNOWLEDGMENTS
Our thanks to College of Science-Mustansiriyah University for supporting this work.
REFERENCES
1. K. Hammood Al-jizani NAA and FSF. Variational Iteration Method for Solving Riccati Matrix Differential
Equations. Indones J Electr Eng Comput Sci. 2017;5, 3:673-683.
2. S. Ghaleb and YQ. New Development of Adomian Decomposition Method for Solving Second Order Ordinary
Differential Equations. Int J Math. 2017;6, 2:28-49.
3. Mohammed DAM and H. The Adomian Decomposition Method for Solving Nonlinear Partial Differential
Equation Using Maple. APM. 2021;11, 6:290-304.
4. T. Tzer Lu and WQZ. Adomian decomposition method for first order PDEs with unprescribed data. Alex Eng J.
2021;60,2:2563-2572.
5. M. kwong, C. Sing, and T. H. A brief Introduction to the Adomian Decomposition Method, with Application in
Astronomy and Astrophysics. Rom Astron J. 2019;1:1-41.
6. M. Richard, W. Zhao. Padé-Sumudu-Adomian Decomposition Method for Nonlinear Schrödinger Equation. J

Appl Math,. 2021;8:1-19.
7. S. Ghaleb YH. New Development of Adomian Decomposition Method for Solving Second Order Ordinary
Differential Equations,. Int J Math,. 6,:28-49,.
8. G. Adomian. Solving Frontier Problems of Physics: The Decomposition Method. Published online 1994.
9. S. Deniz and N. Bildik. Comparison of Adomian decomposition method and Taylor matrix method in solving
different kinds of partial differential equations. IJMO. 2014;4:114-122.
100001-6
10. A. Wazwaz. Analytical solution for the time-dependent Emden-Fowler type of equations by Adomian
decomposition method. Appl MathComput. 166, 3,:638-651,.
11. Lesnic D. A computational algebraic investigation of the decomposition method for time dependent problems.
Appl Math Comput. 2018;119:197-206.
12. H. Bulut and D. Evans. On the solution of the Riccati equation by the decomposition method. Int J Comput
Math. Published online 2002:79, 103-109.
13. Y. Wang LW and WZ. Application of the Adomian decomposition method to fully nonlinear Sine-Gordon
equation,. Int J Nonlinear Sci,. Published online 2006:29-38.
14. M. El-Tawil AB and AA-N. Solving Riccati differential equation using Adomian’s decomposition method.
Appl Math Comput. 157:503-514.
15. T. Lu WZ. Adomian decomposition method for first order PDEs with unprescribed data,. Alex Eng J,.
2021;60,:2563-2572,.
16. Raslan DE and K. Adomian Decomposition Method for Solving Delay Diơerential Equation. Int J Comput
Math. 2004;8:1-6.
17. A. Wazwaz. A Reliable modification of Adomian decomposition method”,. Appl MathCompt,. 1999;102:77-
86.
18. L. wenjin and P. Yaani. Application of Adomian decomposition method to nonlinear systems Adv. Differ Equ.
2020;67:1-17.
100001-7
An efficient Semi Analytic Technique for Solving Non-
Linear Initial Value Problems
Khalid Hammood AL-Jizani
Author Affiliation
Department of Mathematics, College of Science, Mustansiriyah University ,Baghdad, Iraq
Author Email
Abstract. As it known, differential transform approach (DTM) is an analytical technique to solve a different kinds of non-
linear form and usually obtains solutions in series formula. A traditional DTM give a good approximations only in a small
intervals and in the neighborhoods to initials positions but it failed with larger intervals. Therefore, in this work, a reliable
technique of DTM, is proposed which called a multi-step DTM, for increasing the intervals and extended the convergent
region to a series solutions. It deal with approach to obtain accurately the analytic solution for different non-linear problems.
A solution conducted via multi-step DTM is a perfect approximation to more large intervals, as well as locally initially
positions. Mean, in the first and second examples the traditional DTM failed for giving more accurate solution as well as
failed for covering a larger intervals, while multi-step DTM overcome these difficulties successfully. Numerical results
show that this technique is more accurate, reliable and efficient.
INTRODUCTION
Exact solutions of many equations usually is difficult to obtain, especially for nonlinear problems therefore many
approximate analytical methods appeared for this purpose1,2,3. Firstly, DTM is a nice example for finding the solution
analytically and numerically. It was found by Zhou. It was implemented to obtain different kinds of non-linear
fractional, partial, functional, integral and ordinary differential equations. It is based on Taylor series which construct
an analytic solutions in form of polynomials
LITERATURE REVIEW
DTM is approximate analytic approach to solve different types equation. This approach give as a solution in
convergent series terms as well as easy and simple computable components. It suggested by Zhou4 for a nonlinear
initials problem in electric. It give exact value to nth-derivative for the analytical functions to points to term rapidly.
It construct, to equation form with analytic solutions as forms of polynomials. No same a classic higher-order Taylor
series technique which required computations. It’s iteration process to obtain Taylor series solution 5,6. With those
features, solution diverges to several high nonlinear problems that which a boundary condition to infinite7. Many
application for DTM may be presented at 8-29 It introduced a successful technique to varies applications in different
fields of science and engineering. With DTM, it can find a series solution, actually a truncated series solution. These
results doesn’t appears the real demeanor of the problems but give a good approximation to the true solution in a very
small intervals and region. To overcoming that, we suggest reliable technique to DTM, to enlarge and cover the large
interval of the convergence region, it namely a multi-step DTM (MSDTM), which present currently, works a
convergent series solutions over a larger regions as well as improving the accuracy to DTM. Selected examples
verified a validation, accurate and efficiency to this approach.
100002-1
METHODOLOGY
A DTM is good analytic method and give good results for strong linear and nonlinear partial , fractional, functional
and ordinary differential equations. When we need to conduct a solutions for larger interval we observe that the DTM
is fail. To find this analytic solutions with larger intervals and for overcoming this difficulties we use a technique
which is a multi-step method a combination with DTM to find a solutions in larger intervals.
Analysis of DTM
It’s a nice example as an approximate analytic technique to non-linear partial with ordinary formula which use
formal of a polynomial as approximate of the closed form which sufficient differentiable. Idea with definition for
DTM with application for many types of differential equations can be found in8,9,10,11. For convenienceation of the
researcher, we introduce a brief of DTM for the kth derivative of h(t ) function as bellow:
1 ª d s h(t ) º
H ( s) « » , (1)
s ! ¬ dt s ¼ t t0
H ( s ) is a transformed function and h(t ) is the original function. A differential inverse transform for H ( s ) :
f
h(t ) ¦ H (s)(t t ) .
s 0
0
s
(2)
From 1 with 2, we get:

f
(t t0 ) s d s h(t )
h(t ) ¦
s 0
s! dt s t t0
, (3)
Means concept of DTM is derive from Taylor expansions, but a process non estimate a derivative symbolic. However,
relative derivatives which conducted via iteration manner what describe it via transformation equation to originally
functions. To this purpose, a function is described via a finitely series and 2 will be
Z
h(t ) | ¦ H (s)(t t ) .
s 0
0
s
(4)
Z is decided by the convergence of natural frequency. A mainly step to DTM, as a tools to solve many kinds of non-
linear problem.
Multi-Step DTM
Usually the DTM is presented for obtaining approximate solutions to many kinds of non-linear problem in term of
convergence series by simple computation components, but there obstacles, the series solutions always converge in
so smaller regions and it has slow convergent rate or completely divergent when bigger area11,12,13,14. To overcome the
shortcoming and for extended for the convergence region with larger and wider intervals, we introduce a multi-step
DTM (MSDTM). Consider the nonlinear initial value problem:
h(t , q, qc,..., q ( r ) ) 0, (5)
with initial condition q s (0) cs , for s 0,1,..., r 1.
100002-2
Suppose [0, B ] is interval 15, 16, we whish to obtain a solution to IVP (5). In fact implementations for DTM, an analytic
results to IVP (5) have the finite series form as:
Z
g (t ) ¦ b t , t [0, B].
i 0
i
i
(6)
A mult-step technique presents a new idea to construct an approximate solution17,18. Suppose the interval [0, B] divide
into L subintervals [tl 1 , tl ], l 1,2,..., L equal step w B / L with use nodes tl lw. A main idea for MSDTM as
follow. Firstly, implement the DTM to equation (5) over the interval [0, t1 ] , so will conduct a following approximate
solution.
N
g1 (t ) ¦b t ,
i 0
1i
i
t [0, t1 ], (7)
With use the initial condition u1( s ) (0) cs . for l t 2, with all subinterval [tl 1 , tl ] will apply the initial condition
(s) (s)
g l (tl 1 ) g l 1 (tl 1 ) with using equation of DTM (5) over the interval19, 20.
[tl 1 , tl ], And t 0 in equation (1) replaced by tl 1. the process is repeat to generate approximate solution as a sequence
form gl (t ), l 1,2,..., L, for the solutions g (t ).
N
gl (t ) ¦b (t t
i 0
li l 1 )
i
, t [tl , tl 1 ], (8)
Where Z N .L . Actually, the MSDTM suppose solutions as:
g1 (t ), t [0, t1 ]
° g (t ), t [t , t ]
° 2 1 2
°
°.
g (t ) ®
°.
°.
°
°̄ g L (t ), t [t L 1 , t L ].
A new approach, MSDTM 21,22, is easily to compute performance to all values of w. One can simply observe when
step size w B, the MSDTM become a traditional DTM. The main advantages for this approach, it conducted series
solution converges for wide time regions and more extended of the convergence regions 23,24.
TESTING EXPERIMENTS
For explain the effective and efficiency for a propose technique as an approximation tools to solve non-linear problems
of differential equations, the MSDTM is implemented in the following examples25, 26.
Example 4.1
Consider the non-linear IVB yc 1 y 2 (t ), 0 d t d 5,

e 2t 1
with initial condition y (0) 0 , and a closed form as y (t ) .
e 2t 1
100002-3
Now, we apply the MSDTM and comparison with the standard DTM 27, 28, 29. Table 1 explain that.
Table 1. Results for Two approaches
t Exact Solution Absolute Error for DTM Absolute Error for MSDTM
0 0 0 0
0.5 0.462117 4.08741 ×10-9 6.88777 ×10-17
1.0 0.761594 4.19863 ×10-4 3.44222 ×10-17
1.5 0.905148 3.04061 ×10-1 3.61762 ×10-17
2. 0.964028 29.5048 8.90739×10-17
2.5 0.986614 972.087 9.68507×10-17
3. 0.995055 16394.9 6.41723×10-17
3.5 0.998178 175565. 2.35678×10-17
4. 0.999329 1.35433 ×106 3.55525×10-17
4.5 0.999753 8.15391 ×106 1.10106×10-16
5. 0.999909 4.04375 ×107 2.22045 ×10-16
Example 4.2
Consider the non-linear IVB
yc 4 y y3 (t ), 0 d t d 5,
2e4t
with initial condition y (0) 0.5 , and a closed form as y (t ) .
15 e8t
Now, we apply the MSDTM and comparison with the standard DTM. Table 2 explain that.
Table 2 Results for Two approaches
t Exact Solution Absolute Error for DTM Absolute Error for MSDTM
0 0.5 0 0
0.5 1.77141 9.08539 ×10-2 4.64563 ×10-11
1.0 1.99499 1193.49 4.73918 ×10-13
1.5 1.99991 1.11942 ×106 8.92305×10-15
2. 2. 9.92267 ×107 2.89891×10-16
2.5 2. 3.01954 ×109 3.28204×10-16
3. 2. 4.81859 ×1010 2.87761×10-16
3.5 2. 4.96774×1011 3.42645×10-16
4. 2. 3.73172 ×1012 1.62134×10-17
4.5 2. 2.20429 ×1013 2.6109×10-18
5. 2. 1.07794 ×1014 0
From the above examples, more accurate solutions and more extended of the convergence region with larger and wide
intervals.
100002-4
CONCLUSION
For this study, a MSDTM is proposed carefully. An effective modification for the DTM, that get batter the
convergence of the series solution. A proposed technique gives a immediately and evident symbol term of analytics
solution, in addition approximate solutions numerically for non-linear problems. Importantly, one able for overcoming
a restriction for applications areas of standard DTM. Selected examples show the validity of the proposed technique
which has been successfully by getting more extended of the convergence region with larger and wider intervals which
is lead to increase intervals of convergent of series solutions. Other advantage to proposed technique, its applied
directly and not need to make linearization’s with assumption. This emphasizes which fact that DTM is applicable
for different kinds of non-linear case as well as it reliable and promising when compare with an exist approaches.
ACKNOWLEDGEMENTS
Authors would like to express our gratitude to College of Science-Mustansiriyah University for supporting this work.
REFERENCES
1. Khalid Hammood Al-jizani, Noor Atinah Ahmad, and FSF. Variational Iteration Method
for Solving Riccati Matrix Differential Equations. Indones J Electr Eng Comput Sci.
2017;Vol. 5:pp 673-683.
2. Khalid Hammood AL-Jizani, Noor Atinah Ahmad FSF. He’s Variational Iteration Method
for Solving Riccati Matrix Delay Differential Equations of Variable Coefficients,. AIP Conf
Proc. 2017;Vol. 1830,:pp 1-10.
3. Khalid Hammood AL-Jizani NAA. Stability of Linear Delay Differential Equations using
Modified Algebraic Approach. J Telecommun Electron Comput Eng. 2016;Vol. 8, No:pp
157-163.
4. Zhou J. K. Differential Transformation and Its Applications for Electrical Circuits.
Huazhong Univ Press Wuhan, China. Published online 1986.
5. M. Rashidi, and M. Keimanesh. Using Differential Transform Method and the Padé
Approximant for Solving MHD Flow in a Laminar Liquid Film from a Horizontal
Stretching Surface,. Math Probl Eng. 2010:Article ID491319.
6. Rashidi M. and Keimanesh M. AB and HK. Magnetohydrodynamic BioRheological
Transport Phenomena in a Porous Medium: A Simulation of Magnetic Blood Flow Control
and Filtration. Int j numer method biomed eng. 2011;Vol. 27, N:pp 805-821.
doi:doi:10.1002/cnm.1420
7. M. Rashidi. The Modified Differential Transform Method for Solving MHD Boundary-
Layer Equations. Comput Phys Commun. 2009;Vol. 180:2210-2217. doi:10.1016 /
j.cpc.2009.06.029.
8. Fatma Ayaz. Solutions of the system of differential equations by differential transform
method. Appl Math Comput. 2004;147:pp 547–567.
9. Fatma Ayaz. Application of differential transform method to differential-algebraic
equations. Appl Math Comput. 2004;152:649–657.
10. A. Arikoglu, I. Ozkol. Solution of boundary value problems for integro-differential
equations by using differential transform method. Appl Math Comput. 2005;168:pp 1145–
1158.
11. Bildik N., Konuralp A. BF and KF. Solution of different type of the partial differential
equation by differential transform method and Adomian’s decomposition method,. Appl
Math Comput. 2006;127:551–567.
100002-5
12. Arikoglu A. OI. Solution of difference equations by using differential transform method,.
Appl Math Comput. 2006;173 (1):pp 126–136.
13. I AA and O. Solution of differential difference equations by using differential transform
method. Appl Math Comput. 2006;181 (1):153–162.
14. Liu H. and Song Y. Differential transform method applied to high index differential-
algebraic equations. Appl Math Comput. 2007;184 (2):748–753.
15. Momani S. and Noor M. Numerical comparison of methods for solving a special fourth-
order boundary value problem,. Appl Math Lett 191 218–224. Published online 2007.
16. Hassan I. Comparison differential transformation technique with Adomian decomposition
method for linear and nonlinear initial value problems,. Chaos Solitons Fractals 36 53–65.
Published online 2008.
17. Hassan I. Application to differential transformation method for solving systems of
differential equations,. Appl Math Model 32 2552–2559. Published online 2008.
18. El-Shahed M. Application of differential transform method to non-linear oscillatory
systems, Commun. Nonlinear Sci Numer Simul 13 1714–1720. Published online 2008.
19. Odibat Z. Differential transform method for solving Volterra integral equation with
separable kernels,. Math Comput Model 48 144–1149. Published online 2008.
20. Momani S. ZOS and VE. Generalized differential transform method for solving a space-
and time-fractional diffusion-wave equation,. Phys Lett A 370 (5- 6) 379–387. Published
online 2007.
21. Erturk V. MS and OZ. Application of generalized differential transform method to multi-
order fractional differential equations,. Commun Nonlinear SciNumer Simul 13. Published
online 2008:1642–1654.
22. S. OZ and M. Generalized differential transform method for linear partial differential
equations of fractional order,. Appl Math Lett 21. Published online 2008:194–199.
23. S. Momani ZO. A novel method for nonlinear fractional partial differential equations:
Combination of DTM and generalized Taylor’s formula, J. Comput. Appl Math 220.
Published online 2008:pp 85–95.
24. Odibat Z. MS and E V. Generalized differential transform method:Application to
differential equations of fractional order,. Appl Math Comput 197. Published online 2008:pp
467–477.
25. C. KB and Lo. Application of the differential transformation method to the solution of a
damped system with high nonlinearity,. Nonlinear Anal TMA 70. Published online
2009:1732–1737.
26. M. Al-Sawalha MN. Application of the differential transformation method for the solution
of the hyperchaotic Rössler system, Commun. Nonlinear Sci Numer Simul 14. Published
online 2009:1509–1514.
27. C. CS and C. Application of the differential transformation method to the free vibrations of
strongly non-linear oscillators, Nonlinear Anal. RWA 10. Published online 2009:pp 881-
888.
28. K. KA and A. Two-dimensional differential transform method for solving linear and non-
linear Schrödinger equations,. Chaos Solitons Fractals 41. Published online 2009:pp 2277–
2281.
29. Kanth A.and Aruna K. Differential transform method for solving the linear and nonlinear
Klein–Gordon equation,. Comput Phys Commun 180. Published online 2009:pp 708–711.
100002-6
A new Technique of Semi Analytic Approach D Homo for
Obtaining the Analytic Solution to Quadratic Differential
Equations: Theory and Applications
Khalid Hammood AL-Jizani
Author Affiliations
Department of Mathematics, College of Science, Mustansiriyah University ,Baghdad, Iraq
Author Emails
Abstract In this work, a D Homo technique for solving quadratic differential equations is presented by introducing an
auxiliary parameter D , such that D is a small value (generally is less than one) for improving the area of convergence and
getting a wide area for a convergent by iteration process. D Homo as a modification of homotopy approach (MHAM
or D Homotopy ) . Results for this types conducted via a classical homotopy analysis (HAM) grant nice results but
for neighbourhood to initially place. Mainly advantage of a present technique is that it can enlarge and more extended of a
convergent area by iteration process. Results conducted via D Homotopy gives a batter approximations for whole
bigger and wider interval time given, instead of a locally initially place. Some theorems are proved. Numerical results
which obtained, show that this technique is more an accurate, easily and effectiveness.
INTRODUCTION
In this study, D Homo technique is discussed for addressing the quadratic differential equation :
°uc(t ) b(t ) r (t )u (t ) o(t )u 2 (t ),
® 0dt dT (1.1)
°̄u (0) E ,
when b( s), r ( s) and o( s ) are continuous functions, in addition more important in different fields science. For
instance, as is famously, a single-dimensional types are so close quadratic formula. The exact solutions seems to be
difficult conducted especially in nonlinear case, therefore, Many approximate methods recently appeared for this
purpose 1, 2.
LITERATURE REVIEW
At 1992, Liao applied a basic idea related to homotopy topological to present generally approximate analytical
approach to non-linear equations, so called Homotopy approach (HAM). According to homotopy topological. It
generally analytic technique to obtain a series solution for different non-linear differential equations. It’s provide a
simply for ensure a convergent region. It is applied to solve non-linear differential equations 3. A semi analytic HAM
is used to solve fuzzy fractional equation 4. Nematallah 5, obtained Fuzzy impulsive fractional differential equations
obtained by HAM. Saad K. et.al 6, solved a cubic isothermal auto-catalytic chemical system via HAM. Salah A. and
100003-1
Ishak H 7 applied HAM to solve high fraction type using modification of derivation. Saied A. et.al 8, discussed the
convergence and application for HAM on fractional integro-differential equations. Zhong X and Liao 9, implemented
HAM for backward/forward stochastic differential equations. Hariharan G 10, applied HAM for the non-linear partial
differential equations in engineering. More importantly, it provide us the well-known ADM, which a famous
approach is called a homotopy where the auxiliary parameter
11 with H (t ) 1, and when varies, we can
extend the convergence region to achieve an accurate solution 11. Lanre A.12, implemented it for solving seven order
time fraction type Korteweg-devries with a Sawada Kotera form.
A motivation for the proposed technique is to get a highly an accurate solution through expand more extended of
the convergence region and it gives the solution in infinite series. The most important advantages of this technique the
capability provide a continuously formula for approximation solutions, with best details related to results with whole
and wide with larger convergent area of interval time. Other feature, D Homotopy is overhead exist approaches to
solve this type, which a solutions for equation (1.1) conducted by a current technique is efficiency for a bigger values
as well as a small.
METHODOLOGY
Our technique for solving quadratic differential equations by introducing an auxiliary parameter α, such that α is a
small value (generally is less than one) for improving the area of convergence and getting a wide area for a convergent
by iteration process. α - Homo as a modification of homotopy approach (MHAM or α - Homotopy).
Traditional Homotopy Approach
In this section, HAM 13, 14, is applied. A mainly conception, we take the following equation
M [u (W )] 0, (2.1)
M be non-linear, W independently, u (W ) unknown function, respectively. By mean of general of classical HAM,

Liao, construct a zero order form
(1 p) L[I (W ; p) u0 (W )] p H (W ) M [I (W ; P)], (2.2)
p [0,1] as a parameter, z 0 with non-zero as a parameter, H (W ) z 0 a function, L is an auxiliary linearly

operator, u0 (W ) as initially condition of u (W ), I (W ; p ) unknown function, respectively. We can observe
I (W ;0) u0 (W ), I (W ;1) u (W ),
f
then I (W ; p ) u0 (W ) ¦u
m 1
m (W ) p
m
, (2.3)
1 w mI (W ; p)
when um (W ) . ( 2.4)
m ! wp m p 0
f,
when all these correctly selected, one obtain : u (W ) u0 (W ) ¦u
m 1
m (W )
and equation (2.2) be : (1 p) L[I (W ; p) u0 (W )] pM [I (W ; p)] 0,
100003-2
Now, based on equation (2.4), define a vector
u0 {u0 (W ), u1 (W ),..., un (W )}.
then L[um (W ) Fmum1(W )] H (W )Gm (um1)), (2.5)
1 w m1M [I (W ; p)]
when Gm (um1 ) , (2.6)
(m 1)! wp m1 p 0
0, m d 1,
and Fm ®
¯1, m ! 1,
Now, a solution for mth-order form (2.5) after applying the inverse operator to both sides with given initial condition,
for m t 1 becomes:
t
um (t ) F mum1 (t ) ³
0
H (W )G(um1 (W ))))dW , m 1, 2,.... (2.7)
Therefore, easy to solve equation (2.7) by any technique as Mathematic to get an analytic solution
of the HAM. As a famous strong tools, one has:
Theorem (1) (Banach Theorem)
Suppose B Banach space and D: B o B

non-linear mapping, with assume D[u] D[w] d E u w , u , w B
with several constant E 1. Then D has unique fixed point. Also, a sequence um1 D[um ],
with an arbitrary option to u0 B, converge to fixed point D. Based on Th. 1, a non-linear mapping
t
D[u (t )] F mu (t ) ³
0
H (W )G(u (W ))d
))dW (2.8)
A necessary condition for convergence of HAM strictly contraction of D. In addition, equation (2.5) converge to the
fixed point of D, which a solution to equation (2.2).
D Homo Technique to Solve equation (1.1)
A mainly hurdle for original HAM, it a sequence for approximate results conducted quickly converge just foe a
small area. For enlarging and more extended to convergence region for a sequence of approximate conducted, it must
make a modification for HAM via introduce auxiliary parameter (D ) .
(Generally less than one).For equation 2.7, it will become as:
t
um (t ) F mum1 (t ) D ³
0
H (W )G(um1 (W ))dW , m 1, 2,.... (2.9)
100003-3
In fact, a formula (2.9) give a vast freedom to select. Related to complex non-linear equations, it may select a small
value D (in general below one) for getting a best approximations with larger regions. Additionally, we can observe
where D 1, equation (2.9) becomes the standard homotopy analysis formula (2.7).
Theorem (2)
Suppose that u0 (t ) E with {um (t )} conducted by (2.9) converges to u (t ); and u (t ) a solution to equation
(1.1). Proof. Take limit to both sides of equation (2.9), it obtain:
t
lim um (t )
mof mof ³
lim F mum1 (t ) D lim [ H (W )G(um1 (W ))]dW ,
0
mof
(2.10)
t
Then
³
D lim [ H (W )G(um1 (W ))]dW
0
mof
0. (2.11)
t
we know D z 0, we conclude that: ³
0
H (W )G(um1 (W ))
))]dW 0. (2.12)
that’s mean the differential equation for the both sides respect to t yield:
uc(t ) b(t ) r (t )u (t ) o(t )u 2 (t ). ( 2.13)
means u (t ) satisfy equation 1.1 and u (0) E since um (0) E . Then u (t ) is a solution of equation (1.1). Based
on Th.1, easily get a convergence condition to um conducted by (2.9).
Theorem (3)
t
Define a non-linear mapping T [u (t )] F mu(t ) D ³
0
H (W )G(um1 (W ))dW .
A sufficiently condition to converge {um (t )} conducted by (2.9), is strictly contraction for non-linear mapping
T . Additionally, equation (2.9) converge to fixed point of T which it’s a solution to equation (1.1). Related to (2.9),
we select a properly D u0 (t ), then approximate solution for (1.1) at whole interval [0, T ] may be conducted.
TESTING EXPERIMENTS
Currently, we implement MHAM discussed to different types of quadratic differential equations. Result shown
D Homo very simply, more an accurate and effectiveness.
Example 1
°uc(t ) 1 t 2 u 2 (t ),
Consider the quadratic DE ® 0 d t d 4, (2.14)
°̄u (0) 1,
100003-4
e t
2
where the exact solution is u (t ) t t

.
³
1 e z dz
2
0
According to (2.9), choosing α = 0.789, m = 20 we can obtain a best approximations on [0,4] Table 1 show all results
and more clearly from the figure1. It is easily to see when we choose the smaller value of |α| we get a wider for the
convergence region for the sequence 2.9. Additionally, as known the HAM provides us a simply way for adjusting
with control a convergence of solution series via using an auxiliary parameter ћ , and by varying ћ 0.9, can get
extend for the convergence region.
TABLE 1 Comparison Between The Exact And Approximate Solution With ћ 0.9.
t MHAM Absolute Error for D Homo

Exact Solution Absolute error h=-0.9
0.1 1.00031731031038 1.00031731031038 0 5.10703 ×10-15
0.2 1.00241982548044 1.00241982548044 0 1.9984×10-15
0.3 1.00779458802813 1.00779458808213 1.77636 ×10-15 5.9952×10-15
0.4 1.01765087886255 1.01765087886255 1.9984 ×10-15 9.32587×10-15
0.5 1.03295757602714 1.03295757602655 5.85088×10-13 1.42109×10-14
0.6 1.05446681095717 1.05446680992334 1.033383×10-9 1.05249×10-13
0.7 1.0827278201234 1.08272748135379 3.3877×10-7 2.19553×10-10
0.8 1.1181332962412 1.11809254540645 4.07508×10-5 8.78677×10-8
0.9 1.16325470676843 1.16072397273281 2.53073×10-3 1.20523×10-5
1. 1.30926876999662 1.210559901466964 9.86698×10-2 8.24145×10-4
1.08
1.08
1.06
1.06
1.04
1.04
1.02 1.02
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Figure 1 Figure 1
The exact and approximate solutions of x20 (t ) The exact and approximate solutions of x20 (t )
when
1
1 when 0.9
Example 2
° y (t ) y 2 (t ) 1,
Consider the quadratic DE ® 0 d t d 6, (2.15)
°̄ y (0) 0,
e 2t 1
where the exact solution is given by y(t ) .
e 2t 1
100003-5
According to 2.9, choosing D 0.7, m 30, we can obtain a best approximations on [0, 6]. Table 2 show all
results and more clearly from the figure 2. It is easily to see when we choose the smaller value of D , we get a wider
for the convergence region for the sequence 2.9. Additionally, as known the HAM provide a simply way for adjusting
with control a convergence of solution using an auxiliary parameter ћ 0.9 , by varying ћ can get extend for the
convergence region.
Table 2 Comparison between the exact and approximate solution with ћ 0.98
t MHAM Exact Solution Absolute Error for D Homo Absolute error h=-0.98
0.1 0.099667994624955 0.099667994624955 5.55112 ×10-17 2.77556 ×10-17

0.2 0.197375320224905 0.197375320224905 4.71845 ×10-16 2.77556×10-17
0.3 0.291312612451591 0.291312612451591 1.66533 ×10-16 5.5112×10-17
0.4 0.379948962255231 0.379948962255225 5.66214 ×10-15 5.5112×10-17
0.5 0.462117157259964 0.46211715726001 4.62408×10-14 1.66533×10-16
0.6 0.537049566998625 0.537049566998035 5.8975×10-13 2.44249×10-14
0.7 0.604367777111803 0.604367777117164 5.3606×10-12 8.97591×10-12
0.8 0.66403677026172 0.664036770267849 6.12888×10-12 3.89247×10-10
0.9 0.716297870816293 0.716297870199024 6.17269×10-10 1.66493×10-10
1. 0.761594155570312 0.761594155955765 3.85453×10-10 4.57173×10-7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Figure 2 Figure 2
The exact and approximate solutions of x30 (t ) The exact and approximate solutions of x30 (t )
when
1
1 when 0.98
CONCLUSION
In this study, a α - Homo is discussed for solving quadratic form. Comparison with the exact solution is conducted,
result for numerical experiments explain that a proposed technique may grant more an accurate solutions with large
regions and wider interval time, which is a main advantage of the proposed approach. Furthermore, it may overcoming
a restriction for application area for standard HAM. In general, where |α| selected more smaller, we get an accurate
results and more extended of the convergence region for whole time intervals.
ACKNOWLEDGEMENTS
We would like to express our gratitude to College of Science-Mustansiriyah University for supporting this work.
100003-6
REFERENCES
1. K. Hammood Al-Jizani NAA and FSF. Variational Iteration Method for Solving Riccati
Matrix Differential Equations. Indones J Electr Eng Comput Sci. 2017;5(3):673-683.
2. K. Hammood Al-Jizani NAA and FSF. He’s variational iteration method for solving Riccati
matrix delay differential equations of variable coefficients. Proc 4th Int Conf Mthematical
Sci (ICMA4),. 2016;(Palm Garden Hotel, Putrajaya, MALAYSIA, AIP Publishing.).
3. S.Sarath MB and KSS. Fractional Generalised Homotopy Analysis Method for Solving
Nonlinear Fractional Differential Equations. Comput Appl Math. 2020;39(112):1-32.
4. N. Najafi. Homotopy Analysis Method to Solve Fuzzy Impulsive Fractional Differential
Equations. Int J Mod Math Sci. 2020;18(1):58-75.
5. N. Nematallah. Homotopy Analysis Method to Solve Fuzzy Impulsive Fractional
Differential Equations. Int J Mod Math Sci. 2020;18(1):58-75.
6. K. Saad OI and PA. An Effective Homotopy Analysis Method to Solve the Cubic Isothermal
Auto- Catalytic Chemical System,”. AIMS Math. 2018;3(1):183-194.
7. A. Salah HI. Homotopy Decomposition Method for Solving Higher-Order Time Fractional
Diffusion Equation. Sains Malaysiana. 2018;47(11):2899-2905.
8. S. Abbasbandy, M. Hashemi HI. On Convergence of Homotopy Analysis Method and its
Application to Fractional Integro Differential Equations. Quest Math. 2013;36(1):93-105.
9. X. Zhong SL. On the Homotopy Analysis Method for Backword/Forward- Backward
Stochastic Differential Equations. Numer Algorithms. 2017;76(2):487-519.
10. G. Hariharan. On A homotopy Analysis Method for the Nonlinear Partial Differential
Equations A rising in Engineering. Int J Comput Methods Eng Sci Mech. 2017;18(3):191-
200.
11. J. Cang, Y.Tan HX and SL. Series Solutions of Non-linear Riccati Differential Equations
with Fractional Order. Chaos Solitons and Fractals. 2009;40(1):1-9.
12. A. Lanre. q-Homotopy analysis method for solving the seventh-order time-fractional Lax’s
Korteweg-de Vries and Sawada-Kotera Equation. Comput Appl Math. 2019;38(4):187-199.
13. S. Liao. Beyond Perturbation: Introduction to the Homotopy Analysis Method. Chapman
Hall/ CRC Press Boca Rat. Published online 2003.
14. H. Omar. An Integrated Genetic Algorithm and Homotopy Analysis Method to Solve
Nonlinear Equation Systems. Math Probl Eng. Published online 2021:1-14.
100003-7
Chitosan Effect in Chitosan/Hydroxyapatite/Magnesium
Nanocrystalline Composite for Bone Graft Candidate
Fadhil Muhammad Tarmidzi,1, a) Marvel Shaddai Simanungkalit,1) and Nicole Barcelona
Widodo1)
1)
Chemical Engineering Department, Institut Teknologi Kalimantan, Jalan Soekarno-Hatta KM. 15, Balikpapan, Indonesia
a)
Corresponding author: fadhil.tarmidzi@lecturer.itk.ac.id
Abstract. The purpose of this research is to study the effect of chitosan in bone graft composite candidate from a combination of
Hydroxyapatite and Magnesium to its characters such as compressive strength, functional group, and crystallinity.
Hydroxyapatite/Magnesium (HAp/Mg) powder was synthesis via sol-gel method by reacting Ca(NO3) 2.4H2O, MgCl2.6H2O and
(NH4) 2HPO4 in alkaline condition. HAp/Mg was sintered at 600oC for 2 hours. HAp/Mg powder was added to Chitosan (Ch) solution in
malic acid with concentration varies to 5, 7, 9, and 11%w/v and Ch/HAp/Mg ratio 2:1 w/w. IR spectra shows the characteristics peaks of
HAp/Mg composite and interaction between chitosan and HAp/Mg. XRD pattern also shows peaks corresponding to similar reported
HAp/Mg composites. Addition of Chitosan involves in reducing composite crystallite size and compression strength. Maximum
compression strength was achieved at 11%w/v chitosan concentration with 14.45 MPa with crystallite size 23.14 nm. This result was
comparable with the highest strength reported of trabecular bone.
Keywords: Chitosan, Hydroxyapatite, Magnesium, Nanocrystalline composite, Bone graft
INTRODUCTION
Autograft and allograft are technique to treat bone disease such fractures, infections, and tumors. Its limitation, from secondary
operation to probability of infectious diseases become a drawback. Several studies show the use of synthesis graft using alloplast
technique become more reliable resulting the increased demand. One of the main properties needed to developed a synthesis graft
are biocompatibility and osteoconductive.
Within decades, synthetic hydroxyapatite (HAp) has become the most widely used material in biomedical application as bone
grafts due to its component similarity with human hard tissue, bone and teeth1. Due to its outstanding biocompatible and
osteoconductive properties. Synthetic HAp can easily fuse with bone and increasing the bone regeneration process because it has
similar crystallographic and chemical properties of natural HAp. Synthetic HAp may represent a promising bone graft material.
However, HAp has low mechanic strength, which is often combined with other element to increase its strength 2.
Several methods were reported to synthesize HAp from its element include precipitation, sol-gel, and hydrothermal. Sol gel
method is effective method to synthesize HAp due to the ability to control the process. Synthetic HAp via sol gel method also
shown the improvement of contact and stability at the bone interface 3,4.
Magnesium (Mg) and Chitosan (Ch) was added into HAp due to one of the most abundant cations beside calcium. The presence
of Mg is needed to facilitate the growth of osteoblast and osteoclast that can inhibit bone loss (osteoporosis) 5,6. Chitosan also has
been widely used in the medical application as biopolymer which usually combined with bone replacement material. Chitosan
has osteoconductive, biocompatible, biodegradable, and anti-microbial properties which can improve wound healing and
promote osseointegration of bone implant 6.
MATERIALS AND METHODS

Synthesis HAp/Mg
HAp/Mg was synthesized using Ca(NO3)2.4H2O, MgCl2.6H2O and (NH4)2HPO4 obtained from SAP chemical. HAp was
synthesized by diluting a mixture of Ca(NO3)2.4H2O and MgCl2.6H2O in 200 mL aquadest with Mg/Ca theoretical molar
ratio of 0.1. (NH4)2HPO4 was diluted in 200 mL aquadest with Ca/P theoretical molar ratio of 1.67. (NH4)2HPO4 solution
were added drop wise into Mg/Ca solution at 600 rpm to provide a controlled reaction into HAp/Mg. The process was
controlled by adding ammonia solution to develop a base condition at pH 10-11. HAp/Mg solution was aged for 24 hours to
precipitated HAp/Mg. HAp/Mg powder were dried at 150 oC for 8 hours followed by sintering in muffle furnace at 600oC for
2 hours.
Fabrication of Ch/HAp/Mg Nanocrystalline Composite

Chitosan with medical grade 99% deacetylation degree (DD) were diluted in 5% w/v malic acid solution. Chitosan
were varied at 2.5; 3.5; 4.5; and 5.5 gram of chitosan in 50 mL 5% w/v malic acid solution resulting chitosan solution
with concentration of 5, 7, 9, and 11% w/v. Chitosan solution were mixed with HAp/Mg powder with weight ratio of
110001-1
1:2. This mixture was kneaded until well mixed and developed a paste texture resulting Ch/HAp/Mg composite. The
composite was shaped in 10 cm height by 3 cm diameter of acrylic box and rested for 1 hours. Finally, this shaped
composite was dried at 70oC for 24 hours.
Characterization of Ch/HAp/Mg Nanocrystalline Composite
HAp/Mg/Ch composite were analyzed using Universal Testing Machine (UTM) at range 4 – 12 MPa to determine its
compressive strength. Energy Dispersive X-Ray (EDX) were performed to see its elemental composition of
Phosphorus (P), Calcium (Ca), and Magnesium (Mg). Characterization of Ch/HAp/Mg crystallinity and its crystal size
were identified using X-Ray Diffraction Method (XRD). functional groups of composites were identified using
Fourier-Transform Infrared Spectrometer (FTIR).
Effect of Chitosan on HAp/Mg/Ch Nanocrystalline Composite Compressive Strength

Synthesized of HAp/Mg/Ch composite has been successfully performed. Variation of chitosan concentration have a
significance effect on its compressive strength. Addition of chitosan will increase its compressive strength linearly
from 3.61 MPa with 4%w/v chitosan to 14.46 MPa with 11%w/v chitosan (Figure 1). This result proved that the poor
mechanical properties of HAp are supported by the addition of chitosan. This range of compressive strength has
similar to human trabecular bone (Table I) in several sites shows the possibilities of bone graft application.
16 14.46
Compressive Strength (MPa)
14 12.00
12
10 8.43
8
6 3.61
4
2
0
4% 6% 8% 10% 12%
Chitosan Concentration (%w/v)
FIGURE 1. Effect of Chitosan on Composite Compressive Strength
TABLE I. Mean Values of Compressive Strength for Various Anatomic Site7
Anatomic Site Age Range Compressive Strength, MPa

Vertebra 15 – 87 2.4 + 1.6
Proximal tibia 59 – 82 5.3 + 2.9
Proximal femur 58 – 85 6.8 + 4.8
Calcaneus - 1.4 + 1.3
Ch/HAp/Mg Nanocrystalline Composite Elemental Composition

Elemental composition of Ch/HAp/Mg composite were performed to identified the molar ratio of Ca/P. Molar ratio of
Ca/P will affect the synthesis of calcium phosphate compound. Study reported that ratio of 1.67 is needed to developed
HAp (Table II).
TABLE II. Calcium Phosphate Compound and Their Calcium/Phosphorus Ratio8
Compound Formula Ca/P Ratio

Dicalcium Phosphate Anhydrous CaHPO4 1.00
α-Tricalcium Phosphate α-Ca3(PO4)2 1.50
β-Tricalcium Phosphate β-Ca3(PO4)2 1.50
Hydroxyapatite Ca5(PO4)3OH 1.67
Tetracalcium Phosphate Ca4(PO4)2O 2.00
110001-2
The result presented by EDX Analysis (Table III) revealed the quantification of Ca, P, and Mg in %mol. The Ca/P
ratio in Ch/HAp/Mg Composite is 1.59 by mol. By combining Mg concentration as inseparable component in the
composite into the ratio resulting (Ca+Mg)/P mol ratio increase to 1.64. Inclusion of Mg into ratio were reported
because Mg cation can penetrate and substitute with Ca cation to developed HAp/Mg composite 9. The initial molar
ratio of Mg/Ca was 0.1, but EDX analysis shows the final ratio of Mg/Ca only 0.03. This result similar to the latest
reported 0.03 Mg/Ca ratio from magnesium carbonate-free apatite (MCAp) 10 and magnesium-substituted HAp 11.
TABLE III. Quantification for Ca, P, and Mg in Ch/HAp/Mg Composite with 11% Chitosan Concentration
Determined from EDX Analysis
Composition, %mol
Samples P Mg Ca/P (Ca+Mg)/P Mg/Ca
Ca
11% Ch/HAp/Mg 22.97 14.46 0.71 1.59 1.64 0.03
MCAp10 - - - - 1.57 – 1.81 0.03a
Mg/HAp11 - - - - 1.67 – 1.70 0.03a
a
Approximated data from various variable
Effect of Chitosan on Ch/HAp/Mg Nanocrystalline Composite Crytallinity

The formation of a white precipitate during the aging process shows that HAp was successfully synthesized. HAp
precipitate was formed because HAp insoluble in base condition. 24 hours aging process is needed to developed a
crystal due to the saturation concentration in the precipitated. XRD pattern of Ch/HAp/Mg composite shown in Figure
1 was compared and identified with X’Pert Highscore Plus from COD Database No. 96-900-1234. Addition of various
chitosan concentration proves has no effect on its crystallinity. It is known that all the samples exhibit the characteristic
of HAp crystalline corresponded to the database. Individual peaks of Mg have not shown in XRD pattern indicate that
Mg and HAp have a joined structure.
FIGURE 2. XRD Pattern of Ch/HAp/Mg Compared with HAp Database
Crystallinity and its size were calculated by comparing crystalline area with sum of crystalline and amorph area. Table
IV shown that addition of chitosan affects both in crystal size and crystallinity. Increasing chitosan concentration
developed a smaller crystal size and lower crystallinity. All the samples show that Ch/HAp/Mg crystal size can be
classified into nanocrystal ranging from 23 – 30 nm. Lower crystallinity means that the sample have much amorph
phase comparing to apatite crystal phase. Amorph phase is needed in order to develop an elastic-plastic material.
Amorph phase have more elastic properties compared to crystalline phase which have more rigid properties 12,13.
TABLE IV. Effect of Chitosan Concentration on Ch/HAp/Mg Crystal Size and Crystallinity
Chitosan Concentration, %w/v Crystal Size, nm Crystallinity, %

5 30.03 59
7 25.82 57
9 23.02 56
11 23.14 54
110001-3
Effect of Chitosan on Ch/HAp/Mg Nanocrystalline Composite Functional Group
FTIR spectra of Ch/HAp/Mg nanocrystalline composite with variation of chitosan composition are shown in Figure 3.
IR spectral data revealed the characteristic of HAp and chitosan in all the samples. Characteristic of HAp were
observed clearly between 1000 – 1100 cm-1 due to P–O stretching. O–H spectra between 3100 – 3500 cm-1 were hardly
observed either from adsorbed water or vibration from Ch/HAp. The split peak at 565 and 603 cm-1 were identified as
P–O bending of phosphate. XRD pattern shown that there was no present of Mg related crystalline, it also proved in IR
spectra that there were no Mg related peaks at 450 cm-1 14. Peak between 1550 – 1700 cm-1 that related to amide bond
in chitosan were increase with the addition of chitosan concentration 15. The presence of amide bond indicates that the
chitosan in the composite is not degraded.
FIGURE 3. IR Spectra of Ch/HAp/Mg
CONCLUSION
Ch/HAp/Mg nanocrystalline composite was successfully synthesized by sol-gel method followed by sintering at 600oC
for 2 hours. Addition of chitosan in composite shows the increase of its compressive strength. The highest compressive
strength achieved in this research is 14.46 MPa with 11% chitosan concentration. This result has similar strength in
human trabecular bone. IR Spectra also proves chitosan incorporation can increase P – O bonding and no chitosan was
decomposed during sintering process. XRD report shown there was only HAp crystalline developed with crystal size
23.14 – 30.03 nm.
REFERENCES
1
A. Szcześ, L. Hołysz, and E. Chibowski, Adv. Colloid Interface Sci. 249, 321 (2017).
2
M. Prakasam, J. Locs, K. Salma-Ancane, D. Loca, A. Largeteau, and L. Berzina-Cimdina, J. Funct. Biomater. (2015).
3
A.K. Nayak, Int. J. ChemTech Res. 2, 903 (2010).
4
O. Kaygili, S. V Dorozhkin, and S. Keser, Mater. Sci. Eng. C (2014).
5
A. Farzadi, F. Bakhshi, M. Solati-Hashjin, M. Asadi-Eydivand, and N.A.A. Osman, Ceram. Int. 40, 6021 (2014).
6
F. Mammoli, S. Castiglioni, S. Parenti, C. Cappadone, G. Farruggia, S. Iotti, P. Davalli, J.A.M. Maier, A. Grande,
and C. Frassineti, Int. J. Mol. Sci. 20, 1 (2019).
7
T.M. Keaveny, E.F. Morgan, G.L. Niebur, and O.C. Yeh, Annu. Rev. Biomed. Eng. 3, 307 (2001).
8
L.C. Chow, Monogr. Oral Sci. 18, 94 (2001).
9
L. Stipniece, K. Salma-Ancane, D. Jakovlevs, N. Borodajenko, and L. Berzina-Cimdina, Mater. Sci. Appl. Chem. 28,
51 (2013).
10
M.S. Sader, K. Lewis, G.A. Soares, and R.Z. LeGeros, Mater. Res. 16, 779 (2013).
11
I.V. Fadeeva, L.I. Shvorenva, S.M. Barinov, and V.P. Orlovskii, Inorg. Mater. 39, 947 (2003).
12
R. Glüge, H. Altenbach, I. Kolesov, N. Mahmood, M. Beiner, and R. Androsch, Polymer (Guildf). 160, (2018).
13
A. Zima, Spectrochim. Acta Part A Mol. … (2018).
14
V. Karthikeyan, S. Dhanapandian, and C. Manoharan, Int. Lett. Chem. Phys. Astron. 70, 33 (2016).
15
L. Pighinelli and M. Kucharska, Carbohydr. Polym. 93, 256 (2013).
110001-4
Design and Implementation of Vehicle Detection System
Using You Only Look Once and Inverse Perspective
Mapping
Okky Putra Barus1, a) and Irvin Delano1, b)
Author Affiliations
1
Universitas Pelita Harapan
Jl. Imam Bonjol No.6, Medan, 20112, Indonesia
Author Emails
a)
Corresponding author: okky.barus@uph.edu
b)
id70007@student.uph.edu
Abstract. Based on data from the Indonesian National Police regarding the distribution of the factors causing accidents
in Indonesia in 2017, the majority of the causes of traffic accidents in Indonesia are human error factors (amounting to
61%) where every hour, on average, there are 3 new victims of car accidents. The power of AI can be harnessed to reduce
the number of accidents because AI can work with a lower error rate than humans and is more consistent too. In this
study, a vehicle detection system will be built that allows autonomous cars to find out the location of surrounding
vehicles and also make estimates of the relative distance to other vehicles. This vehicle detection system is built using
You Only Look Once and Inverse Perspective Mapping algorithms. This vehicle detection system can process 1 input
image in an average of 3.562 seconds, can process video input data with a sampling rate of 5 FPS as fast as 0.83625 FPS,
and has a distance estimation deviation of ± 0.422 meters.
INTRODUCTION
Based on data from the Indonesian National Police regarding the distribution of the factors causing accidents in
Indonesia in 2017, the majority of the causes of traffic accidents in Indonesia are human factors (61%), followed by
infrastructure and environmental factors (30%) and vehicle factors (9%). Every hour, on average there are 3 new
victims due to traffic accidents in Indonesia. This figure will continue to increase every year in proportion to the
increasing volume of vehicles as well. The power of AI can be used to reduce the number of these accidents. One of
the other advantages of AI is that it can work with a much lower error rate than humans and is more precise. The
advantages of this AI foster an idea that is Self-Driving Car (Autonomous Car). Autonomous cars are cars that can
move independently and safely without human interaction/intervention. Autonomous cars can be the future of the
world's transportation system because they offer safety and efficiency in transportation. With autonomous cars,
accident rates can be reduced.
In this writing, the author will discuss the implementation of the Vehicle Detection System. With this Vehicle
Detection system module, autonomous cars will be able to locate other vehicles and estimate the distance to these
other vehicles. This research focuses more on implementation in combining two methods, namely You Only Look
Once (YOLO) version 3 to perform object detection and Inverse Perspective Mapping (IPM) to perform relative
distance estimation.
120001-1
METHOD
Building Yolo Neural Network Model
The first stage (even before the input image is entered into the system) is the preparation of the model that will
be used to identify the car object. The construction of a neural network model requires a trained network architecture
and matrix weights. The network architecture is built using Keras modules running on top of Tensorflow. The
network architecture was created following the specifications of “YOLOv3: An Incremental Improvement”. The
tensor weights used are from the Darknet team who have been trained with the MSCOCO dataset. This neural
network model can predict as many as 80 different object classes, but in this study, only 1 object classes were used,
namely the car class to predict road images and videos. Then, the Keras library will be used to build the YOLO
neural network architecture. After the neural network architecture has been built, the weight tensor will be loaded
into the neural network architecture, resulting in a neural network model in the format of HDF5 (Hierarchical Data
Format 5) format.
Inverse Perspective Mapping Calibration
So that IPM can know the mapping that must be done to the input image, it must be calibrated first. In this study,
calibration was carried out using a rectangular object (having 4 known point coordinates). After IPM already knows
the 3 x 3 transformation matrix H, then the input image can be converted into a bird's eye view version. From the
rectangular calibration tool, 4 corner points of the newspaper in the input image will be obtained. The four corner
points of the newspaper will be recorded in a format (top-left, top-right, bottom-right, bottom-left) which will then
be referred to as calibration points. This transformation matrix will remain constant for a certain location and
rotation of the camera placement, so only 1 calibration is needed. The mapping of coordinates between 2 planes is
define as the following.
Where the relationship above maps the relationship between the pixels (x,y) of the bird's eye view image and the
pixels (u,v) of the input image. The 3 x 3 dimension matrix in the above dimensions is called the H matrix which
must be calculated for to perform Inverse Perspective Mapping. To calculate the value of the H matrix, we need 4
coordinate points that are already known to be located in the input image and also in the bird's eye view image later
(a total of 8 coordinate points).
Input Image Specification and Preprocessing

The input image will first be resized to the dimensions of 416px x 416px (width x height) according to the shape
of the input layer of the YOLO model. The input image will be parsed into its pixel tensor form. The input image
pixel tensor has dimensions H height x W width x D depth where height and width indicate the dimensions of the
photo and depth indicates the 3 available color channels, namely Red, Green, and Blue (RGB). The input image
pixel tensor still cannot be used directly. The pixel tensor must be normalized first into the range 0 – 1 (inclusive) by
dividing all elements by a scalar value of 255. The value of 255 was chosen because considering the value of 1 pixel
unit in the pixel tensor has a range of 0 - 255 (inclusive) which is the range of image RGB channel values. The
normalization process also applies to input data with video format. The difference is that the normalization process
is carried out on each frame that is sampled from the video input data. In addition, all normalization processes are
the same for image and video input data.
Bounding Box Identification by Neural Network Model
The normalized pixel tensor will be entered as input (feed forward) into the neural network that has been
prepared in the previous stage. The Keras library will preload the HDF5 file containing the YOLO neural network
model. Then, Keras Library will be used to do 1 time feed forward with the input pixel tensor. Next, the neural
network will generate an output tensor with dimensions 13 × 13 × [3 (4 + 1 + 80)] (small scale), 26 x 26 × [3 ‫( כ‬4 +
120001-2
1 + 80)] (medium scale), and 52 × 52 × [3 (4 + 1 + 80)] (large scale), where [3 (4 + 1 + 80)] means 3 bounding
boxes, 4 bounding offsets information (in the form of x coordinates of the bounding center point box, the y-
coordinate of the middle point of the bounding box, the width of the bounding box, and the height of the bounding
box), the probability of an object (objectness score), and the 80 probability of the predicted object class.
Unfortunately, among the bounding boxes generated by predictions from the YOLO neural network, there is still a
lot of overlap. In fact, this problem is very common in other object detection algorithms. To overcome these
overlapping bounding boxes, Non-maximum Suppression will be applied.
Filtering based on Confidence Score & Elimination of Overlapping Bounding Boxes
Each bounding box has a confidence score (range 0-1, inclusive) which is determined by the neural network.
Thus, we need an optimal threshold or limit value (usually > 0.6) so that it can minimize false positives when
filtering out the many bounding boxes generated by the neural network. After filtration, there are still many
overlapping bounding boxes (there are more than 1 bounding box for 1 object). Therefore, Non-maximum
Suppression is carried out which aims to eliminate excessive bounding boxes that overlap with the main bounding
box of an object. Through Non-maximum Suppression, it can be guaranteed that for every detected object there is
only 1 bounding box that marks it. The reason the filtration stage is carried out before Non-maximum Suppression is
because the Non-maximum Suppression variant used has an exponential complexity of N^2 (N to the power of 2)
where N is a lot of data. By doing the filtration stage first, the less data that must be processed by Non-maximum
Suppression.
Road Perspective Mapping with IPM

The input image is still in the camera perspective, so other car objects are also still visible through the camera
perspective. From a camera perspective, the distance between one car and another car is still not 1:1 because the car
that is close will look bigger, while the car that is far looks smaller. To be able to estimate the distance of other cars,
a 1:1 scale representation is needed, so the Inverse Perspective Mapping method is used to map the input image to
its 1:1 representation. The coordinates of the bounding box will be transformed first using IPM and then all the
coordinates of the bounding box will already be linear in distance, therefore the distance can be estimated easily.
IPM requires calibration points in order to know how to map an image from the camera perspective to its 1:1
representation. Therefore, a newspaper as a calibration tool that has been expanded into 2 pages is used. The four
corners of the newsprint (top-left, top-right, bottom-right, and bottom-left) will be used as calibration points because
these four corners will definitely form a square/rectangle in the 1:1 representation later.
FIGURE 1. Comparison of original image (left) and transformed image (right)
Estimating Relative Distance

In this research simulation, the camera used to capture the input image is mounted on the windshield of the car,
causing a blind spot due to obstruction by the car hood as illustrated in Figure 1. Therefore, the distance estimation
result from the system will be added again with the car hood blind spot distance to produce the final estimation
result. After the bounding box has been transformed with IPM, the distance between the blind spot of the hood and
the point below the center of each bounding box will be estimated using the Euclidean Distance method which can
be used to measure the distance between 2 points. To estimate the distance, it is necessary to provide information on
the ratio of the actual distance (in meters) per 1 pixel in a 1:1 image representation. In the simulation later, the
known width of the calibration tool in the form of a newspaper that has been opened into 2 pages will be used as a
comparison to get the meter per pixel value so that the distance from other vehicles can be estimated.
120001-3
MAIN RESULTS
In this experiment, the system is used to detect the black SUV car and uses a calibration tool in the form of a
newspaper that has been opened into 2 pages as can be seen in Figure 2.
FIGURE 2. Highlight of calibration tool (newspaper)
By using a tool made by the author, it is found that in the input image, the newspaper object is at the coordinates:
a. Coordinates of the top-left corner point : 652, 942.6666870117188
b. Coordinates of the top-right corner point : 817, 937.666687017188
c. Coordinates of the bottom-right corner point : 821, 969.6666870117188
d. Coordinates of the bottom-left corner point : 637, 97566666870117188
Then, the four newspaper objects will be mapped to a new plane with dimensions of 3500px x 3500px. Because
the newspaper object is definitely a rectangle, the four points of the newspaper in the new plane can be determined
as follows:
a. Coordinates of the top-left corner point : 1650, 1650
b. Coordinates of the top-right corner point : 1850, 1650
c. Coordinates of the bottom-right corner point : 1850, 1850
d. Coordinates of the bottom-left corner point : 1850, 1850
Based on the information from the calibration tool, an H transformation matrix can be obtained which can be
used to map the coordinates of each bounding box to its Bird's Eye View perspective. Here is the value of the matrix
H obtained.
[
[ -0.54489462 -2.71819437 2211.5886583 ]
[ -0.11402304 -5.41513127 4472.97296225]
[ -0.00001524 -0.00150421 1. ]
]
It is also known that the length of a newspaper that has been opened into 2 pages is 0.81 meters, so by doing the
IPM transformation first, it is found that the meter value per pixel is 0.00404999752807681 meter per pixel. Final
output from the system can be seen on Figure 3. The final output contains the drawn bounding boxes of each object,
detected object class label, and the respective estimated relative distance in meter.
FIGURE 3. Final result of ML pipeline
120001-4
Below is a list of execution duration trials for 10 ML pipeline trials in Google Collaboratory. Experiments were
carried out with GPU-enabled settings. Based on the table below, the average execution time is 3.562 seconds for 1
image.
TABLE 1. Image data processing execution list
Execution No. Execution Time (second)
1 3,65
2 3,53
3 3,46
4 3,49
5 3,47
6 3,78
7 3,47
8 3,47
9 3,77
10 3,53
Average 3,562
It can be seen in table 1 that the execution duration by the system is relatively slow compared to the YOLO
performance that should be (usually less than 1 second) when the YOLO algorithm is applied to the image dataset.
This is because this execution experiment was carried out on a Google Collaboratory which has a shared GPU
instance, so this experiment cannot utilize the GPU's power 100%.
Below is a list of execution durations for 8 trials in Google Collaboratory. Experiments were carried out with
GPU-enabled settings and with a sampling rate of every 5 FPS. Based on the table below, the average obtained is
0.83625 FPS. This FPS value is low because the resources in Google Collaboratory are shared.
TABLE 2. Video data processing execution list
Execution No. FPS
1 0,89
2 0,88
3 0,83
4 0,81
5 0,82
6 0,83
7 0,82
8 0,81
Average 0,83625
It can be seen in table 2, that the execution duration by the system is relatively slow compared to the YOLO
performance that should be (usually more than or equal to 30 FPS) if the YOLO algorithm is applied to the video
dataset. This is because this execution experiment was carried out on a Google Collaboratory which has a shared
GPU instance, so this experiment cannot utilize the GPU's power 100%.
The following is a list of distance estimation experiments. In the table below, the actual distance is compared
with the estimated distance from the system. The actual distance used has been adjusted by subtracting the hood's
blind spot measured at 2.7 meters.
TABLE 3. List of distance estimation experiments
Real Distance from Bumper
Execution No. Real Distance (meter) Estimated Distance (meter) Deviation (meter)
(meter)
1 11,00 8,30 10,13 1,83
2 4,20 1,50 1,62 0,12
3 4,35 1,65 1,72 0,07
4 4,85 2,15 2,23 0,08
5 4,95 2,25 2,24 0,01
Deviation Average 0,422
Based on the experimental list table above, the average deviation of the estimated distance is ± 0.422 meters. In
other words, the result of the estimated distance from the system may be larger or smaller by 0.422 meters from the
actual distance.
120001-5
CONCLUSIONS
From the results of the implementation and experimental development of the Vehicle Detection System with
YOLO and IPM, some points can be concluded. The average duration of processing execution time for 1 (one)
image using Google Collaboratory with GPU-enabled settings is 3.562 seconds. This duration is relatively slow
(should be at least < 1 second) because the experiment is run on a shared instance (hardware that is used by several
users at the same time). The average duration of processing execution time for a video with a sampling rate of 5 FPS
using Google Collaboratory with GPU-enabled settings is 0.83625 FPS. This FPS value is relatively low (should be
at least 30 FPS) because the experiment was run on a shared instance (hardware used by several users at the same
time). The average deviation from the estimated distance is 0.422 meters, which means that in a distance estimation
measurement, the distance estimation results can be 0.422 meters larger or smaller than the actual distance.
This research is still far from perfect, especially with the many obstacles found, such as the difficulty of
obtaining a dataset of the distance from other cars on the actual highway and the unknown camera skew & distortion
factor. Therefore, so that the implementation results can be even better, several things that can be done for further
research. It is necessary to take into account the distortion factor of the camera lens, so that the accuracy of distance
estimation can be improved. The need for additional training (transfer learning) on the YOLO model to the
Indonesian car dataset, because sometimes there are bounding boxes that do not cover the car perfectly (too small or
too big). The need to find a complete dataset with data on the distance between cars if any, so that research results
can be evaluated more accurately.
ACKNOWLEDGEMENT
This work is supported by LPPM Universitas Pelita Harapan.
REFERENCES
1. Chris. (2020, Febuari 9). What is padding in a neural network? Retrieved from Machine Curve:
https://www.machinecurve.com/index.php/2020/02/07/what-is-padding-in-a-neural-network
2. Escott, E. (2017, Oktober 24). What are the 3 types of AI? A guide to narrow, general, and super artificial
intelligence. Retrieved from Codebots: https://codebots.com/artificial-intelligence/the-3-types-of-ai-is-the-
third-even-possible
3. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. 1, 2-3, 6.
4. Himanshu. (2019, Januari 19). Activation Functions : Sigmoid, ReLU, Leaky ReLU and Softmax basics for
Neural Networks and Deep Learning. Retrieved from Medium: https://medium.com/@himanshuxd/activation-
functions-sigmoid-relu-leaky-relu-and-softmax-basics-for-neural-networks-and-deep-8d9c70eed91e
5. Jordan, M. I., & Mitchell, T. M. (2015). Machine learning: Trends, perspectives, and prospects. Washington,
D. C.: American Association for the Advancement of Science.
6. Kathuria, A. (2018, April 23). What’s new in YOLO v3? Retrieved from Medium:
https://towardsdatascience.com/yolo-v3-object-detection-53fb7d3bfe6b
7. Muljono, Sutojo, & Andono, P. N. (2017). Pengolahan Citra Digital. Yogyakarta: Penerbit Andi.
8. Nurhikmat, T. (2018). Implementasi Deep Learning Untuk Image Classification Menggunakan Algoritma
Convolutional Neural Network (CNN) pada Citra Wayang Golek. 4-5 ,21, 22-23, 31.
9. Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. 2.
10. Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2015). You Only Look Once: Unified, Real-Time Object
Detection. 1-2, 4-5.
11. Rena, P. N. (2019). Penerapan Metode Convolutional Neural Network pada Pendeteksi Gambar Notasi Balok.
11.
12. Shi, Z. (2019). Advanced Artificial Intelligence. Beijing: Chinese Academy of Sciences.
13. Singh, A. (2020, August 3). Selecting the Right Bounding Box Using Non-Max Suppression (with
implementation). Retrieved from Analytics Vidhya: https://www.analyticsvidhya.com/blog/2020/08/selecting-
the-right-bounding-box-using-non-max-suppression-with-implementation/
14. Tan, D. (2020, May 14). A Hands-On Application of Homography: IPM. Retrieved from Towards Data
Science: https://towardsdatascience.com/a-hands-on-application-of-homography-ipm-18d9e47c152f
120001-6
COVID-19, SME Training Support by ASEAN
Governments and Analysis of Digital Investments in Brunei
Shahid Anjum 1, a) and Nurul Syahidah Binti Haji Sharin 1)
Author Affiliations
1
Universiti Teknologi Brunei, Tungku Link Road, Gadong
Brunei Darussalam, BE 1410
Author Emails
a)
Corresponding author: anjumsw@hotmail.com
a)
shah.jum@utb.edu.bn
Abstract. Digital adoption and related investments among SMEs can be challenging during COVID-19 smart business
adjustments due to their small size and limited capabilities especially in ASEAN region where contribution of SMEs is
over 99%. In order to overcome this barrier, it is necessary for every country’s government to provide SMEs with the right
support that can help them in their digital transformation journey. The negative impact of the pandemic phenomenon has
been huge on adaption and employability of human resources and SME’s technological innovation globally. The support
of ASEAN governments in the context of human resources development dimension in digital space is reviewed in this
article. Besides, digitalization of Brunei’s SMEs has been lukewarm although the public of the country has been an active
user of social media where the subscription rate of mobile usage is around 136% and internet usage around 94%. Therefore,
in order to provide the evidence of explanatory theory on the digital technologies’ adoption among SMEs in Brunei, this
study has used both qualitative and quantities approaches, has collected data from 101 Brunei SMEs through survey
questionnaire and analyzed it using partial least square with the help ADANCO. The study has found that the internal
capacity of organization in the form of digital skills and external capacity in the form of digital infrastructure in the economy
are statistically significant in explaining the importance of digital educational support provided by the government towards
the adoption of digital technologies. The challenges faced by SMEs due to pandemic related restrictions in the form of
decrease in sales, access to finance, budget constraints for employee salaries and wages, procurement and order delivery
related difficulties, however, have not statistically necessitated the adoption of digital technologies because of policy
incentives for digital technology related education and training facilitations.
INTRODUCTION
With the current situation of the outbreak of COVID-19 pandemic that causes international public health concern,
it has had drastic impact on the economics of every country globally. Small enterprises are particularly vulnerable
towards the effects of the governments’ measure to contain the virus due to its small assets and capability to sustain
themselves. The pandemic has different effects on different industry as some are needed to be open and some of
forcibly shut down (Bartik et al. 2020). There are also other factors that may cause these differences such as low
capability or lack of the right resources. This is mostly due to the negative impact of the phenomenon on SME’s
technological innovation and human resources such as delayed of projects, delayed of new product launch, and delayed
of recruitment (Zou et al., 2020). According to technology experts, this global disruption creates either a challenges
or opportunities for businesses to improve their business processes by transforming their business models or
implement new technology (Gregurec et al., 2021). If SMEs wanted to boost their performance, they should consider
engaging in digital transformation (Soon et al., 2019). Even though there are significant increase of investment of
technology and employee skills development after the pandemic as compared to previous years, implementation of
digital adoption comes with high rate of failed results (Chutijirawong et al. 2019). There are also insufficient digital
130001-1
skills such as data scientist and social marketers become one of the majority challenges in digital adoption among
businesses in ASEAN countries (Loh et al., 2020).
ASEAN had been investing their efforts on economic growth. ASEAN’s GDP growth of 7% is the result of the
digital economy and integration of digital could contribute USD1 Trillion towards ASEAN’s GDP in 2025 (Hoppe,
et al., 2018). According to ASEAN’s official website, SMEs had contributed significantly to the labour force of AMS
and contributed 30% to 53% of ASEAN GDP. There are various efforts by ASEAN to ensure the development of
SMEs including the establishment of agencies to formulate policies, programms and activities on the development of
SMEs. The paper will be focused on SMEs digitization efforts to in Brunei as well as ASEAN Member States (AMS)
supports to ensure SMEs in ASEAN region are able to sustain themselves during the difficulty times of COVID-19
pandemic. Besides, this study has collected data from SMEs through questionnaire to provide explanatory theory on
the digital technologies’ adoption among SMEs in Brunei. The paper starts with section 1 is introduction, section 2 is
on review of literature and section 3 describes data and empirical estimation results. Section 4 concludes the paper.
REVIEW OF LITERATURE
There are various drivers of change in business models of SMEs in during the pandemic such as social changes,
customer centricity, organisational capabilities, financial aspects, competitor trends, market trends and other
environmental trends (Gregurec et al, 2021). However, in terms of digital transformation during the pandemic, the
most common drivers is social changes and followed by customer and organisational drivers. This is mainly due to
the various measure taken by government across the globe such as lockdowns and restriction that could changes in
consumer’s purchasing behaviour, forced the organisation to adopt quickly towards the digital-oriented environments
such as acquiring knowledge, utilizing relevant digital technologies, and learn new skills. In effort to supports digital
transformation across ASEAN especially during the pandemic, ASEAN launched ASEAN Digital Masterplan 2025
(ADM 2025) that aims to promote digital economy and digital society. This masterplan has been developed to
overcome current global issues of recovering from the pandemic, climate change and technology trends. ADM
identified eight desirable outcomes that should be achieved by 2025 through enabling actions. One of the desired
outcomes of the plan is developing a digitally inclusive society in ASEAN, which concerns on the long-term adoption
of digital technologies among the ASEAN citizens. According to research, lack of digital literacy among end-users
are the top barriers from achieving the ADM 2025 Vision. Thus, this outcome focused on four identified barriers of
digital adoption across ASEAN citizens, which are accessibility, affordability, motivation, and skills where skills of
SME entrepreneurs are an integral part. This outcome concern on the long-term adoption of digital technologies among
the ASEAN citizens. According to research, lack of digital literacy among end-users are the top barriers from
achieving the ADM 2025 Vision. Some of the actions in ADM 2025 are based on the continuing actions and efforts
from ADM 2015. This includes initiatives to ensure that citizens have access to internet and the skills to use digital
technologies, which has become even more important because of pandemic. It has been reported that 64% of SMEs
are adopting digital technology to enhance their customer and end-user experience such as through utilizing data to
obtained customer insights. In response to the pandemic, firms had taken initiatives to increase their efforts in
technologies and innovation. More than half of firms had changed their operations remotely including their workplace,
recruitment, business negotiation through digitalization (Zou et al., 2020).
Even small businesses experiences some sort of digital transformation due to the pandemic There are also some
firms that use technologies as their marketing strategies such as livestreaming, social media platform, social marketing
and so on. Ultimately, different organizations adopt technologies at different level according to their digital
capabilities, resources, target market and business needs (Chen et al., 2021). This is also considering that the more
affected the sector are, the more adjustment that they need to made. Most common technologies that attracts most
researchers are social media and platforms with 32.9% appearances in the publications and mobile technologies with
28.2% appearances (Gregurec et al., 2021). SMEs investment on technology consists of 76% are on software and
services such as digital marketing and website. Some of investment include 57% on ICT hardware and network and
52% on ICT training programmes for employee. It also identified that those technology invested are targeted for
different area: 74% for digital sales and services, 70% for digital marketing and 41% for operational processes (Loh
et al, 2020). Due to SMEs contribution to a nation’s economy, it is essential for governments to provide necessary
support for their small businesses to sustain especially during these tough times. To assist SMEs, supports can come
in two form, which are short-term and long term. Long-term measures refer to incentives to ensure that their SMEs
are able to cope and sustain in the post-COVID 19 environment. (OECD, 2020) which must definitely include the
digital training aspects of SMEs and entrepreneurs (Chen et al., 2021).
130001-2
TABLE 1: COVID-19 related Human Resources Digital Incentives by ASEAN government
Country Online Trainings/ Courses Access to experts
Hands-on training by MPEC,
partner up with IT companies to Access to learning platform
Online classes accredited by experts globally provide programmes such as by global accredited experts
Brunei
(OECD 2020a; Hayat M, 2021; MOFE 2020) programming, analytic data and (OECD 2020a; APEC, 2020;
other IR 4.0 technologies (MOFE Intarakumnerd et al., 2020)
2020)
Mentorship are also available
Funds for boot camp for reskilling
through the partnership
workers (UNCTAD, 2020).
Educational program that allows female start- (UNDP, 2020). Get advice
Cambodia Partner-up to provide training
ups to flourish (Heuisler, 2020) and support from technical
through incubation, acceleration
experts and business coaches
(UNDP, 2020)
(Chan, 2020)
Pilot online classes for least developed areas.
(Sakudo, 2021) competences such as cloud Conducts training for SMEs (Loh
Indonesia N/A
computing, programming, developer (OECD, et al., 2020)
2020b)
Lao PDR N/A N/A N/A
Short courses in digital skills and
Online entrepreneurial platform (Dayangku,
Malaysia highly skilled courses (Cusmano, N/A
2020)
2020; Sagar, 2020)
Training courses through e-learning platforms
Myanmar N/A N/A
(swisscontact, 2020)
E-Learning platform: online courses,
interactive seminars, and in-depth training on
Training program by Google PH Programme for foreign
disaster risk reduction, business continuity,
to provide MSMEs knowledge specialists to transfer
Philippines and climate change adaptation.
and digital skills to run their capabilities to local workers
E-learnings to participate in training,
businesses. (OECD., 2020b)
workshops, seminars and advices from
international experts (PDRF, 2020).
Educational workshops for local CTO- that enables SMEs to
business owners and community connect with digital
Free online course provider Coursera (Chee,
Singapore leaders to pick up digital skills to consultants with expertise IT
2021)
help them engage customers on consultancy firms (Today
social media. (Chee, 2021) 2021)
Tools and mentoring (Wang J.,
Thailand Program to learn digital skills (Wang, 2020) N/A
2020)
Program that allows
businesses to connect with
Implementation of webinars to boost knowledge
Hybrid Workshop to Support the experts, digital consultancy,
Vietnam of advertising for SMEs (Vietnam Investment
Digitalization (Myint, 2021) solution providers and
Review, 2021)
government agencies
(USAID, 2021)
ASEAN’s ICT workers are only less than 0.5% of the population where most advanced ICT economies are at 2%.
Furthermore, the report also cited that only selected and few primary and secondary schools implement digital
education programme. To achieved successful digital integration, it is very crucial to ensure that the digital skills of
ASEAN population are matched with the current technology by not just focusing on primary and secondary schools
but comprehensive digital education programmes at all level (Hoppe et al., 2018). Government should help SMEs in
improving the managers or owners of small business’s digital capabilities by providing digital training or education
to increase their knowledge on digital technologies that enables them to develop a digital vision in their adoption
process of digital technologies. Digital learning and training systems that are easily access by employees of small
business should also be consider to help small business in saving their money on employing new employees for digital
skills and training (Chen et al, 2021). Government should offer free online learning that could assist SMEs. Brunei
(OECD 2020a; Hayat M, 2021; MOFE 2020) and Singapore (Chee, 2021) offered free online course and classes from
Coursera which are accredited by experts globally. In Philippine, the government had set up e-learning platform that
130001-3
offers self-paced online courses, interactive seminars, and in-depth training on disaster risk reduction, business
continuity, and climate change adaptation (PDRF, 2020). This also being executed by other AMS which are
Malaysia (Dayangku, 2020), Cambodia (Heuisler, 2020), Thailand (Wang, 2020) and Myanmar (Swisscontact,
2020). In Indonesia, the government also pilot online classes for least developed areas (Sakudo, 2021) to boost their
competences such as cloud computing, programming and developing (OECD, 2020b). Meanwhile, Vietnam
implementation of webinars to boost knowledge of advertising for SMEs (Vietnam Investment Review, 2021).
Compare to online courses and classes that can be access at any time, training refers to courses that requires hands-on
and expertise.
FIGURE 1. Graphic Representation of the model (ADANCO results)
Majority of the ASEAN countries (80%) are offering their enterprises with access to training including Malaysia
(Cusmano, 2020; Sagar, 2020), Indonesia (Loh et al., 2020) and Vietnam (Myint, 2021). In Singapore, an educational
workshop are held for local business owners and community leaders to pick up digital skills to help them engage
customers on social media. (Chee, 2021). In Brunei, MPEC (Manpower Planning and Employment Council) had
partner up with IT companies to assist the government to provide hands-on training such as programming, analytic
data and other IR 4.0 technologies (MOFE, 2020). In Thailand, the training include access to tools and mentoring
(Wang, 2020). Cambodia provide supports for this area by providing funds for boot camp for reskilling workers
(UNCTAD, 2020) and partner-up to provide training through incubation, acceleration and mentorship (UNDP,
2020). Lastly, access to expertise could also aid SMEs in gaining knowledge through sharing of experiences
from expertise and only half of the AMS provide access of this support for their SMEs. This includes SMEs in
Cambodia that are able get advice and support from technical experts and business coaches (Chan, 2020) as well as
in SMEs in Vietnam where they could enter a program that allows businesses to connect with experts, digital
consultancy, solution providers and government agencies (USAID, 2021). In Singapore, the government had develop
CTO- that enables SMEs to connect with digital consultants with expertise IT consultancy firms (Today 2021). Then,
Brunei (OECD 2020a; APEC, 2020; Intarakumnerd et al., 2020) and Philippines (OECD, 2020b) offered access to
their SMEs with accredited foreign specialist from all over the whole to share and transfer their capabilities to the
local workers.
130001-4
DATA AND EMPIRICAL ESTIMATION
Data has been collected via a survey questionnaire, distributed through online, from 101 enterprises in Brunei and
has applied partial least square using ADANCO for analysis, a choice made after reviewing the applications of various
competitive techniques like linear regression (Ismail et al., 2021), stepwise linear function (Anjum et al. 2021),
logistic regression (Anjum, 2020), advanced econometric analysis (Shamim et al., 2015 & 2013), regressioneering
and regressgineering (Anjum 2014a & 2014b), efficiency analysis model (Shamim et al., 2017) and data mining
(Anjum 2014c, 2013) etc. The hypothesis is the investment in the digital technologies is one of the key ways to survive
during and in post COVID-19 which public sector need to facilitate. Public administration can provide facilitation
through the understanding of challenges faced by SMEs during the pandemic as well as in digital transformation
journey, drivers towards digital transformation and supports needed from the government in order to feel ease in the
journey of digital transformation. These five dimensions has been approached with constructs having various items.
TABLE 2: Assessments for evaluation of measurement model

Predefined Reliability = 1
Composite reliability Convergent validity
Type of outer model # of items
ρA ρC αC AVE
EDU 3 0.9049 0.938 0.9001 0.8348
LIC 7 0.9099 0.926 0.9063 0.6421
LEC 3 0.8129 0.89 0.8127 0.7301
FTP 4 0.7405 0.754 0.641 0.4664
DSS 3 0.8167 0.873 0.7889 0.6967
DM 5 0.8573 0.8378 0.7792 0.5193
OP 4 0.7836 0.8453 0.7585 0.5804
BE 5 0.7717 0.8234 0.7265 0.4964
CE 3 0.8588 0.9086 0.8496 0.7684
IBP 6 0.8227 0.8575 0.7992 0.5069
Notes: Dijkstra–Henseler’s = ρA; Jöreskog’s rho = ρC; Cronbach’s alpha = αC; Average Variance Extracted = AVE;
Type of outer model: All Latent Variables = All Latent Variables; Predefined Reliability = 1 (For All)
The reliability score results of reflective measurement model, shown in Figure 1, for construct and discriminant
validities are significant based on various criteria (Table 2). Results for the former reliability are good like the ρA
values are above 0.7, all the ρC values are in the range of satisfactory good, values of the internal consistency reliability
measure of αC are above 0.7 (except for FTP), values for AVE show sufficient uni-dimensionality because they are all
above 50%. The results for assessment of discriminant validity as are shown by the values of Heterotrait-Monotrait
Ratio of Correlations (HTMT) and Fornell-Larcker Criterion are all significant except for few ones like OP to FTP,
CE and IBP to BE and LEC to LIC for both and for OP to DSS and DM in case of HTMT only.
TABLE 3: Structural Model Evaluation Assessment

Relation PC f2 Relation PC f2 Relation PC f2
EDU → FTP 0.244 0.086 EDU → DM 0.33 0.175 LIC → EDU 0.581 0.155
EDU → DSS 0.431 0.269 EDU → OP 0.383 0.217 LEC → EDU 0.130 0.008
Notes: Path Coefficient (PC) = Beta; Effect size = Cohen f2 (or f2)
The values for the assessment of structural model are also very reportable. Adjusted R-square values for EDU,
FTP, DSS, DM, OP and BE are all in the range of above 30% to 70%. Moreover, the f2 values ranging from 0.008 to
0.086 and from 0.155 to 0.296 i.e. from weak to medium effect sizes respectively (Table 3).
CONCLUSION
It is found that the digital education (EDU) provided as a measure of promoting digitization of the SMEs plays a key
role in explaining the amount of investment made by the SMEs towards sustainability through pandemic years is
explained through the following estimated model. The investment related main categories considered are digital
payments or fintech (FTP), digital sales and security (DSS), organizational processes and reengineering (OP) and
digital marketing (DM) which receives a boost from variety of adoptive strategy including business expansion (BE),
130001-5
customer engagement (CE) and improving business processes (IBP), besides digital education. Limited internal (LIC)
and external (LEC) capabilities generates a use case for the digital education.
REFERENCES
1. A. Bartik, M. Bertand, Z. Cullen, E. Glaeser, M. Luca and C. Stanton, How are small businesses adjusting to
COVID-19? 2020. DOI: 10.2139/ssrn.3574741
2. Chutijirawong N., Hora V., Bunsupaporn K., Bunyalug C., Satiyathiwat S. The Thailand Digital Transformation
Survey Report 2020, Deloitte, 2019
3. D. N. Z. Ismail and S. Anjum, Privacy Policy, Training and Adaption of EMT to Curtail Workplace Harassment
in Organizations: An Application of TAM”, Advances in Intelligent Systems and Computing Series. Springer
4. F. Hoppe, T. May and J. Lin, Advancing Towards ASEAN Digital Integration. Bain & Company. 2018
5. F. Shamim, N. Yamori and S. Anjum, Clicks Business of Deposit Taking Institutions: An Efficiency Analysis.
Journal of Economic Studies. 44(6). Emerald Publishing. Thomson Reuters, 2017.
6. F. Shamim, S. Anjum and A. B. Wakil, Banking Risk and Operating Efficiency Measures in the Era of IT.
Accounting and Finance Research. Sciedu Press. 4 (1), 2015.
7. F. Shamim and S. Anjum, Technology Diffusion in Japanese Finance Industry: An Exploration. International
Research Journal of Applied Finance, VI (12), (December 2013).
8. I. Gregurec, M. Tomiˇcić Furjan and K. Tomiˇcić-Pupek, The Impact of COVID-19 on Sustainable Business
Models in SMEs. Sustainability 2021, 13, 1098, 2021
9. J. Myint, (2021) US-ABC Holds Hybrid Workshop to Support the Digitalization of Vietnam SMEs. US-ASEAN
Business Council.
10. J. Wang, Being there for Thailand’s small businesses. Thailand. (August 27, 2020)
11. K. Chee, New initiative to offer free digital marketing training for students, professionals and SMEs, The Straits
Times, Singapore, (2021, April 1)
12. L. Chen, and F. Kimura, E-Commerce Connectivity in ASEAN. Economic Research Institute for ASEAN and
EAST ASIA, (2020).
13. L. Cusmano and S. Raes (2020): Coronavirus (COVID-19): SME policy responses. Tackling coronavirus (covid-
19): contributing to a global effort. OECD. p121.
14. L. Loh, G. Descamps and A. Chia, Are you ready to turn today’s challenges into opportunities?: ASEAN SME
Transformation Study 2020, Accenture, 2020
15. L. Soon and C. Chuan, Redesigning for the digital economy: A study of SMEs in Southeast Asia, EY, 2019
16. M. Hayat, Most Read of 2020: Brunei Darussalam: An Unexpected COVID-19 Success Story. Australian
Institute of International Affairs. (2021, January 07)
17. MOFE. Ministry of Finance and Economy. Brunei Darussalam. 2020
18. M. Sagar, Malaysia announces RM20 billion economic stimulus package amid COVID-19. Open Gov. (2020)
19. M. Sakudo, The new normal: digitalization of MSEMs in Indonesia. APEC-Canada Growing Business
Partnership, (2021, February 9).
20. OECD, Enterprise policy responses to covid-19 in ASEAN: Measures to boost MSME resilience, Canada, 2020a
21. OECD Policy options to support digitalization of business models during covid-1: annex. Saudi Arabia 2020b.
22. PDRF, Digital transformation key to recovery of education and business sector amid pandemic. Philippine
Disaster Resilience Foundation Philippine, (November 9, 2020)
23. P. Intarakumnerd and U. Korwatanasakul, Business impacts and government support for firms’ upgrading and
innovation under COVID-19 in Southeast Asia. (2020).
24. P. Zou, D. Huo, and M. Li, The impact of the COVID-19 pandemic on firms: a survey in Guangdong Province,
China. Global Health Research and Policy, 2020
25. S. Anjum, Statistical Software and Regression Diagnostic Reporting with Fuzzy-AHP Intelligent Zax", Lecture
Notes in Software Engineering, Vol. 2, No. 1, February, (2014a).
26. S. Anjum, Back testing VaR Violations, Be-ALAM Regression and Internal Models of Portfolio Variance
Forecasting,” Journal of Money, Investment and Banking, FRDN Inc., 29, Oct. (2014b).
27. S. Anjum, Composite Indicators for Data Mining: A New Framework for Assessment of Prediction Classifiers.
Journal of Economics, Business & Management, 2(1), (Feb. 2014c).
28. S. Anjum, Algorithms for Predictive Classification in Data Mining: A Comparison of Evaluation Methodologies.
Journal of Industrial and Intelligent Information (JIII), 1(2), (June 2013).
130001-6
29. S. Anjum, and N. Qaseem, Basel violations, volatility model variants and value at risk: Optimization of
performance deviations in banks. Economics and Business Letters 10(3), 240-248, 2021.
30. S. Anjum, DoS Attacks, Triad and Privacy: Software Exposures in Microsoft, Apple and Google, IEEE Xplore,
Proceedings of 2020 First International Conference of SMARTTECH, Riyadh, IEEE, Washington, 2020.
31. Swisscontact.org, Hospitality training continues during Covid-19 downtime. 2020
32. UNCTAD (2020) Cambodia’s digital startups help blunt economic impact of covid-19. Switzerland.
33. UNDP (2020, May 12) A new partnership to enhance Cambodia’s youth entrepreneurship and SMEs. Cambodia.
34. USAID.org, USAID helps small and medium enterprises join Vietnam’s digital transformation. (Jan. 29, 2021)
130001-7
System Analysis in Virtual Student Assignments at
University
Jamal1,a), Riyayatsyah1,b), Tri Sudinugraha1,c), Richki Hardi1,d), Hanafi2,e)
Author Affiliations
1
Departement of Informatics, Universitas Mulia, Balikpapan, Indonesia
2
Departement of Informatics, Universitas Amikom Yogyakarta, Yogyakarta, Indonesia
Author Emails
a)
jamal@universitasmulia.ac.id
b)
riyayatsyah@universitasmulia.ac.id
c)
tri.sudinugraha@universitasmulia.ac.id
d)
e)
hanafi@ amikom.ac.id
Abstract. Today the computer is one of the technologies that can increase the effectiveness in the teaching and learning
process. Students can access various kinds of information virtually through computers connected to the internet. The
widespread use of the internet is a potential for learning with an online system. Learning with online or virtual systems
allows students to access information and assignments flexibly without being limited by time and place. Online learning
has many synonyms that are often used, including e-learning, internet learning, distributed learning, networked learning,
del e-learning, virtual learning, computer-assisted learning, Web-based learning, distance learning. All these terms refer
to the distance between students and educators or instructors, where students use computer technology to access subject
matter, interact with educators and other students, and obtain some form of assistance available to students. The purpose
of this study is to provide the results of system analysis in student assignments virtually at the University.
INTRODUCTION
The Covid-19 pandemic has forced the world community to define the meaning of life, the purpose of learning,
and humanity's nature. If all this time, humans are forced to live in a fast-paced situation, work without stopping,
and pursuit of economic growth targets in a competitive system. However, the spread of the Coronavirus, which has
become a significant crisis for modern humans, forces us to breathe for a moment, stop from the vortex of the
system, and look back at life, family, and the social environment in its truest sense1. Humans are forced to eliminate
from their routines, to interpret what is sought from life2.
Indonesia has significant challenges in handling Covid-19. Of all the aspects that become a challenge today,
which concentrates on the educational part, essential for discussion. The education aspect has become the author's
concentration because he has spent decades in this field as a researcher-practitioner to the policymaker.
The Covid-19 pandemic has forced social distancing policies, or in Indonesia, it is better known as physical
distancing to minimize the spread of Covid-19. So, this policy is being pursued to slow down the spread of the
Coronavirus in the community. The Ministry of Education and Culture responded with a study from home policy
through online learning followed by the abolition of the National Examination for this year.
All affected countries have tried to make their best policies to maintain education services' sustainability.
Indonesia also faces several real challenges that must be immediately resolved: (1) technological disparities between
schools in big cities and regions, (2) limited teacher competence in the use of learning applications, (3) limited
140001-1
resources for the use of educational technology such as internet and quotas, (4) the relationship between teacher-
student-parents in online learning is not yet integral.
The implementation of the physical distancing policy, which later became the basis for implementing learning
from home, with the sudden use of information technology, often surprises educators and students, including parents
and even everyone in the house. Information technology learning has indeed been implemented in the last few years
in the education system in Indonesia. However, the online learning that took place as a surprise from the Covid-19
pandemic shocked almost all lines, from districts/cities, provinces, centres and even the international community3.
So, these obstacles are an important note for our education world, which must catch up with online learning
quickly. In fact, not all technology and systems are ready. So far, online learning is only a concept, as a technical
tool, not as a way of thinking, as a learning paradigm. Online learning is not a method to change face-to-face
learning with digital applications, nor does it burden students with piles of assignments every day. Online learning
should encourage students to be creative, access as many sources of knowledge as possible, produce works, hone
insights and ultimately shape students into lifelong learners4.
The Covid-19 pandemic has been a shock to all of us. The world seemed to slow down and even stopped for a
moment. Large and modern countries have been hit hard by the rapid spread of the Corona Virus, resulting in
thousands of deaths spread across various countries. Indonesia has faced many challenges from Covid-19, which
makes us all have to take care of each other together. The five essential issues above will determine how quickly we
will be able to flatten the anxiety curve of students, teachers, principals, parents, and all 5.
Amid the Covid-19 pandemic, our education system must be ready to leap to transform online learning for all
students and all teachers. Entering a new era to build creativity, hone student skills, and improve self-quality by
changing our interaction systems, perspectives, and patterns with technology6.
The online learning system (in the network) is a learning system without face to face directly between teachers
and students but is carried out online using the internet network. Teachers must ensure that teaching and learning
activities continue, even though students are home. The solution is that teachers must be able to design learning
media as innovations by utilizing online media (online).
The learning system is carried out through a personal computer (PC) or laptop connected to an internet network
connection. Teachers can do learning together simultaneously using groups on social media such as WhatsApp,
Telegram, Instagram, zoom applications, or other teaching media. Thus, the teacher can ensure that students take
part in learning simultaneously, even though in different places.
All sectors are feeling the impact of the corona. The world of education is one of them. Judging from the
surrounding events, both students and parents who do not have cellphones to support online learning activities feel
confused, so the school is looking for solutions to anticipate this. Some students who don't have cellphones study in
groups, so they also participate in learning activities. Start learning through video calls connected to the teacher in
question, being asked questions one by one, and taking attendance through Voice Notes available on WhatsApp. The
materials are also provided in videos that are less than two minutes long.
Online learning cannot be separated from the internet network. An Internet network connection is one of the
obstacles faced by students whose residence is difficult to access the internet, especially those students who live in
rural, remote and disadvantaged areas. Even if someone uses a cellular network, sometimes the network is unstable
because the geographical location is still far from the cellular signal coverage. This problem often occurs in students
who take part in online learning, so its implementation is not optimal.
Many on various social media share parents' experiences while accompanying their children to study, both
positive and negative. For example, it turns out that there are parents who are often angry because their children are
unruly, so they can't stand it and want their children to go back to school.
This incident showed parents that educating their children was not easy, requiring a lot of knowledge and
patience. So with this incident, parents must be aware and know how to guide their children in learning. After
getting this experience, it is hoped that parents will learn how to educate their children at home.
It should be realized that the unpreparedness of teachers and students towards online learning is also a problem.
The shift from conventional learning systems to online systems is sudden, without proper preparation. But all of this
must be carried out so that the learning process can run smoothly and students actively participate even in the
conditions of the Covid-19 pandemic.
The stuttering of online learning is evident in front of us, not just in one or two schools but throughout several
regions in Indonesia. Essential components of the online learning process need to be improved and improved. First
and foremost is a stable internet network, a capable device or computer, an application with a user-friendly platform,
and online socialization that is efficient, effective, continuous, and integrative to all education stakeholders.
140001-2
The solution to this problem is that the government must provide a policy by opening free online application
services in collaboration with internet and application providers to assist this online learning process. The
government must also prepare an online-based learning curriculum and syllabus. For schools, it is necessary to
conduct online technical guidance on the online implementation process and conduct socialization to parents and
students through print and social media regarding the procedures for implementing online learning concerning their
roles and duties.
The researchers conducted a system analysis in student assignments completed virtually at universities from this
background.
LITERATUR REVIEW
In online learning, where students and educators are separated, educators must know the learning principles and
how students learn 7. Delivery tools are not a determining factor for learning quality; instead, the design of the
subject determines the effectiveness of learning. In designing an online learning material, no one theory must be
followed but must use a combination of views. The material developer must know learning approaches to choose the
right learning strategy8. Learning methods should be selected to motivate students, facilitate the learning process
from whole people, serve individual differences, promote meaningful learning, encourage interaction, provide
feedback, facilitate contextual learning, and encourage during the learning process9.
Learning theory in online learning.
At first, the learning system using computers was designed based on a behavioristic approach, which states that
the behaviour that can be seen shows whether students have learned something or not and not what happens to
students' minds. In response, some educators stated that not all learning could be seen, and there is more to it than
just behavioural changes. Then came the cognitive learning theory. Cognitive psychology says that education
includes memory, motivation and thinking, and reflection is essential in education10.
METHODOLOGY
The data analysis method in this study is inductive qualitative research, where the data starts from the current
phenomenon and presents the findings. This type of phenomenological research specializes in wonders and visible
realities to examine their explanations, which aims to find meaning or meaning from experiences that exist in life.
DISCUSSION
Based on the research objectives mentioned earlier, this study explicitly examines the application of e-learning
from a pedagogical aspect11. The pedagogical aspect that becomes the reference in researching the application of e-
learning refers to the analysis, design, development, and implementation. Before the lecturers held online courses,
they argued that not all materials could be online12. There are material analysis activities carried out. This activity is
carried out by identifying approaches whose topics are appropriate to be implemented through e-learning13. he
supporting lecturer, the development team, and colleagues carried out the material analysis. The analysis is also
carried out by considering the variety of knowledge of the material. The process is by studying the learning
objectives and then considering whether these learning objectives can be achieved using the concept of e-learning14.
The material analysis results obtained several courses that are suitable to be developed through e-learning. The
material produced is also very dependent on the lecturer. Media analysis and the e-learning environment are also
concerns for lecturers at the University15.
The development team consisting of department heads, web managers, and lecturers who are members of the
development team conducted a study by comparing various platforms for e-learning. The choice of venue is seen
from the one with the highest learning content16. The tools available on the platform are examined only to the extent
of trying to use them and looking at the possibilities for use in learning, such as forums, blogs, assignments, exams,
and online discussions17. The e-learning development team analyzes infrastructure, such as servers, access points,
and networks, in terms of environmental analysis. The analysis has also been carried out by lecturers at the
University, namely the consideration of e-learning learning methods18. This is based on the opinion of the lecturer
that in determining the online and face-to-face learning methods, it must also be relevant to the characteristics of the
students19. The implementation of e-learning also needs to pay attention to the competencies that students must
140001-3
possess in lectures. Therefore, all lecturers at the University believe that determining the type of assignment must
also be relevant to the characteristics of the students.
Lecturers in the department believe that they can give assignments both individually and in groups during online
lectures20. Lecturers feel that in the implementation of e-learning, additional costs are needed to pay for e-learning
support facilities such as internet networks, computers, etc. because in its performance it is not as easy as thought,
lecturers who use e-learning learning strategies must be given incentives that different from lecturers who conduct
face-to-face lectures only21. The implementation of e-learning will have feasibility according to needs, and it is
necessary to make a policy supported by the tertiary level. There is also a need for policy support on e-learning at
the university level. However, during the implementation of e-learning, very few lecturers at the university believe
that they receive support from central policies and campus policies in the form of recognition or appreciation of
learning innovations with e-learning. This is not in line with the opinion of lecturers who feel that e-learning in
universities has become a necessity in this era of globalization. High commitment because it is recognized that e-
learning is very challenging learning. In the implementation of e-learning, lecturers must create exciting learning
content. Even though there is no direct face-to-face meeting, students are still happy to learn lecture material22. The
following results show that students and lecturers are satisfied if the task is carried out virtually with an adequate
system23.
FIGURE 1. Level of satisfaction in virtual assignments at the University
CONCLUSION
A variety of learning activities must be used to accommodate different learning styles. Learners will choose the
right strategy to meet their learning needs. A variety of pre-learning activities can be used to prepare students for the
lesson's details and keep them connected and motivated to learn online lessons. Learners should be informed of the
learning outcomes of the study so that they know what they can expect and can interpret when to achieve that lesson
outcome. Learners should be informed of the requirements as prerequisites to check whether they are ready for the
lesson. Online learners should be provided with various learning activities to achieve learning outcomes and
accommodate individual learner needs. When learners carry out learning activities, they must be involved in
multiple interactions. Learners need to interact with interfaces to access online materials. Based on the results of
data analysis conducted in this study, there are several conclusions, namely e-learning at the university held by
lecturers for courses that can be accessed virtually. Online learning strategies are designed to vary, namely
presenting material, deepening material through online discussion forums, applying knowledge through online
assignments, and measuring learning success through online tests, while face-to-face learning is more emphasized
on discussions, student presentations, demonstrations, and practice. At the development stage, lecturers develop
teaching materials in the form of texts, presentation slides, videos, animations, homemade or utilizing existing media
on the internet, develop their online assignment materials, online discussion forums and online test questions; In
140001-4
terms of e-learning implementation, most lecturers still have difficulty in facilitating virtual classes, especially in
uploading materials, discussion forum assignments and online tests.
ACKNOWLEDGMENTS
This research was supported by the Institute for research development and community service at Mulia
University, Balikpapan City, East Kalimantan, Indonesia. The researchers thank the academic community of Mulia
University have helped researchers in preparing data and documents related to the research.
REFERENCES
1. S. Hambali, A. Akbaruddin, D. Bustomi, A. Rifai, T. Iskandar, A.F. Ridlo, Y. Meirizal, R. Rusmana, and R.A.
Tyas, Int. J. Hum. Mov. Sport. Sci. 9, (2021).
2. M.Z. Hoq, Am. J. Educ. Res. 8, (2020).
3. R. Jhaveri, Clin. Ther. 42, (2020).
4. M. Nicola, Z. Alsafi, C. Sohrabi, A. Kerwan, A. Al-Jabir, C. Iosifidis, M. Agha, and R. Agha, The Socio-
Economic Implications of the Coronavirus Pandemic (COVID-19): A Review 78, (2020).
5. P.A. Vasishta, A.K. Anjali, M.P. Brundha, and V. Sivaswamy, Int. J. Res. Pharm. Sci. 11, (2020).
6. B. Pfefferbaum and C.S. North, N. Engl. J. Med. 383, (2020).
7. O.B. Adedoyin and E. Soykan, Covid-19 Pandemic and Online Learning: The Challenges and Opportunities
(2020).
8. H.C. Wei and C. Chou, Distance Educ. 41, (2020).
9. S.C.H. Hoi, D. Sahoo, J. Lu, and P. Zhao, Neurocomputing 459, (2021).
10. K. Mukhtar, K. Javed, M. Arooj, and A. Sethi, Pakistan J. Med. Sci. 36, (2020).
11. R.A. Carter, M. Rice, S. Yang, and H.A. Jackson, Inf. Learn. Sci. 121, (2020).
12. Mailizar, A. Almanthari, S. Maulina, and S. Bruce, Eurasia J. Math. Sci. Technol. Educ. 16, (2020).
13. M.A. Almaiah, A. Al-Khasawneh, and A. Althunibat, Educ. Inf. Technol. 25, (2020).
15. A.Y. Alqahtani and A.A. Rajkhan, Educ. Sci. 10, (2020).
16. D. Al-Fraihat, M. Joy, R. Masa’deh, and J. Sinclair, Comput. Human Behav. 102, (2020).
18. M.M. Alam, N. Ahmad, Q.N. Naveed, A. Patel, M. Abohashrh, and M.A. Khaleel, Sustain. 13, (2021).
19. K. Regmi and L. Jones, A Systematic Review of the Factors - Enablers and Barriers - Affecting e-Learning in
Health Sciences Education 20, (2020).
20. A. Shahzad, R. Hassan, A.Y. Aremu, A. Hussain, and R.N. Lodhi, Qual. Quant. 55, (2021).
21. N.M. Alqudah, H.M. Jammal, O. Saleh, Y. Khader, N. Obeidat, and J. Alqudah, Ann. Med. Surg. 59, (2020).
22. A. Muniasamy and A. Alasiry, Int. J. Emerg. Technol. Learn. 15, (2020).
23. S. Choi, H.M. Yuen, R. Annan, M. Monroy-Valle, T. Pickup, N.E.L. Aduku, A. Pulman, C.E. Portillo
Sermeño, A.A. Jackson, and A. Ashworth, Arch. Dis. Child. 105, (2020).
140001-5

Wa0004.

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wa0004.

Uploaded by

Copyright:

Available Formats

scitation.

AIP Conference Proceedings Volume 2658

The 2nd International

Balikpapan, Indonesia • 20–21 October 2021

The 2nd International

All papers have been peer reviewed.

Melville, New York, 2022

© 2022 AIP Publishing LLC

No claim is made to original U.S. Government works.

Preface: 2nd International Conference of Science and Information Technology in Smart

PLENARY AND KEYNOTE PAPERS

Android based remote driving for real time vehicle monitoring

The use of e-learning to increase student innovation in technopreneurship

Design of Google Apps-based helpdesk ticketing system at University using extreme

Employee monitoring, productivity, privacy and harassment: Auditing technology training

Improve educational marketing strategy through use of digital marketing technology

Utilization of social media application in regency SMEs performance

An international tourist behaviour on mobile smartphone usage

Sentiment analysis of government regulations regarding the implementation

SCIENCE TECHNOLOGY FOR SOCIETY

Universitas Mulia, Balikpapan, Indonesia

BACKGROUND AND MOTIVATION

Software Security, Cyber Risk Management and Remote Working

Tool Usage Example

FIGURE 3. Homepage system and instruction process to use this system

FIGURE 5. Display result and level of software security readiness

Road Safety and Road Accident

Causal Loop Diagram

FIGURE 1. Step by step to design CLD 19,22,23

FIGURE 2. Road Safety Causal Loop Diagram Solution for Surabaya

Sub Model Endogenous Variable Exogenous Variable

FIGURE 1. shows an activity diagram of the security framework .

The structure of the kchart research is depicted in Figure 2:

FIGURE 3. Interaction stages

Keywords— sentiment analysis, regulation sentiment, social media analytics

Data Collection Dataset Pre-processing Classification Model Analysis of

FIGURE 1. Research Stage

B. DATA SET PRE-PROCESSING

TABLE 1. Example Manual Labeling

TABLE 2. Text Preprocessing Result

TABLE 3. Comparison of Performance

TABLE 5. Feature Ablation

FIGURE 1. Live Streaming Workflow (Kaltura Knowledge Center, 2020)

Smartphone Front Rear Memory

System architecture in this project can be explained as follows:

RESULT AND DISCCUSSION

TABLE 2. Average delay

Agustinus Noertjahyana1, a) Lily Puspa Dewi 1, b) Justinus Andjarwirawan 1, c) and

Fig. 1. Channel Selection Diagram

Fig 3. Main Program

Fig 4. Broadcaster Role

FPS Publish Live Streaming

Inte Christinawati Bu’ulolo1,a), Melani Isabella Siregar1, Clara Fellysa

1. Category All Words are not Equal

1. Jaro Winkler Distance Algorithm

C. Word Similarity Detection

1. Jaro Winkler Distance Algorithm

D. Get the Best Algorithm

RESULT AND DISCUSSION

1. Test results and calculation of the Jaro Winkler Distance Algorithm

2. Test results and calculation of the Winnowing Algorithm

Data MS Word Feature Results Label Winnowing Algorithm Results Label

Learning Communication Media