Page 1 of 5 - Cover Page Submission ID trn:oid:::1:2793903907

Document W
Document
Quick Submit

Quick Submit

Paideia School

Document Details

Submission ID

trn:oid:::1:2793903907 3 Pages

Submission Date 1,653 Words

Jan 3, 2024, 3:20 PM EST

9,601 Characters

Download Date

Jan 3, 2024, 3:20 PM EST

File Name

Document_8.docx

File Size

14.5 KB

Page 1 of 5 - Cover Page Submission ID trn:oid:::1:2793903907

 Page 2 of 5 - AI Writing Overview Submission ID trn:oid:::1:2793903907

How much of this submission has been generated by AI?

38%
Caution: Percentage may not indicate academic misconduct. Review required.

It is essential to understand the limitations of AI detection before making decisions

about a student's work. We encourage you to learn more about Turnitin's AI detection
capabilities before using the tool.
of qualifying text in this submission has been determined to be
generated by AI.

Frequently Asked Questions

What does the percentage mean?

The percentage shown in the AI writing detection indicator and in the AI writing report is the amount of qualifying text within the
submission that Turnitin's AI writing detection model determines was generated by AI.

Our testing has found that there is a higher incidence of false positives when the percentage is less than 20. In order to reduce the
likelihood of misinterpretation, the AI indicator will display an asterisk for percentages less than 20 to call attention to the fact that
the score is less reliable.

However, the final decision on whether any misconduct has occurred rests with the reviewer/instructor. They should use the
percentage as a means to start a formative conversation with their student and/or use it to examine the submitted assignment in
greater detail according to their school's policies.

How does Turnitin's indicator address false positives?

Our model only processes qualifying text in the form of long-form writing. Long-form writing means individual sentences contained in paragraphs that make up a
longer piece of written work, such as an essay, a dissertation, or an article, etc. Qualifying text that has been determined to be AI-generated will be highlighted blue
on the submission text.

Non-qualifying text, such as bullet points, annotated bibliographies, etc., will not be processed and can create disparity between the submission highlights and the
percentage shown.

What does 'qualifying text' mean?

Sometimes false positives (incorrectly flagging human-written text as AI-generated), can include lists without a lot of structural variation, text that literally repeats
itself, or text that has been paraphrased without developing new ideas. If our indicator shows a higher amount of AI writing in such text, we advise you to take that
into consideration when looking at the percentage indicated.

In a longer document with a mix of authentic writing and AI generated text, it can be difficult to exactly determine where the AI writing begins and original writing
ends, but our model should give you a reliable guide to start conversations with the submitting student.

Disclaimer
Our AI writing assessment is designed to help educators identify text that might be prepared by a generative AI tool. Our AI writing assessment may not always be accurate (it may misidentify
both human and AI-generated text) so it should not be used as the sole basis for adverse actions against a student. It takes further scrutiny and human judgment in conjunction with an
organization's application of its specific academic policies to determine whether any academic misconduct has occurred.

Page 2 of 5 - AI Writing Overview Submission ID trn:oid:::1:2793903907

 Page 3 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907

Automatic Summarization of Indonesian Text Using Line Ranking and Decision Tree
Meherzadi Muntaha email: meherzadimuntaha52@gmail.com current position: department of
information technology at the the University of Southern Punjab in Lahore, Pakistan
Summary –This study introduces a novel approach to automatically summarising news articles from
Indonesia. Combining an choice tree trained on fifty example summaries with sentence scoring—
which ranks sentences according to their relevance—makes it work. Using this method, the
algorithm can pick out the most crucial sentences and provide brief summaries that cover all the
bases. With an average accuracy of 0.58 as well as higher peaks of 0.80, the results are encouraging.
This shows that the strategy could be useful for readers since it saves them time and energy without
sacrificing the information they need from news items.
Information Retrieval: text summarising, sentence evaluation, decision tree;
I. INTRODUCTION
Information overload is a problem, especially with Indonesian news articles. It suggests a way to
automatically summarise texts by utilising decision trees and sentence score. It is challenging to
effectively comprehend pertinent news information due to information overload. There is a pressing
demand for reliable summarising techniques since reading summaries saves time compared to
reading whole articles.
Accuracy values below 55% have been found in previous study on Indonesian text summarising.
Dealing with nebulous data presents challenges for traditional summarising methods.In this
investigation, C4.5 decision trees are used in conjunction with sentence scoring. Sentence scoring
uses characteristics like phrase position, term frequency, and keyword frequency to identify
important sentences. Based on the scoring, the C4.5 algorithm chooses the more relevant sentences
to include in the summary. In comparison to earlier studies, this method seeks to increase precision.
Additionally, it uses decision trees, which are efficient, to choose sentences faster. A combination of
recall, precision, and f-measure metrics will be used to assess the summarization system's
correctness.
Taken together, the findings of this study offer a viable strategy for the automated summary of news
articles from Indonesia. It seeks to improve accuracy and processing speeds by integrating decision
trees with sentence scoring..
II. LITERATURE REVIEW
A. Automatic Text Summarization
Condensing a lengthy document into a more concise one while maintaining its essential ideas is the
goal of text summarization. A relevant summary can be automatically generated by picking out the
most crucial elements of a piece of writing or document. There are two primary methods: extraction,
which involves copying key sections directly, and abstraction, which involves creating new sentences
to express the key points. The optimum way to choose or combine extraction and abstraction
approaches for best outcomes in diverse situations is still uncertain, even if they are effective for
summarising..)
B. Related Researches
In 1958, Luhn proposed a method for automated text summarization based on phrase frequency;
later, Baxen-dale improved upon it by adding sentence order as a consideration. In 1969, Edmunson
expanded on this by adding other aspects such as title and key words. Diverse approaches have
flourished in the field since then. The TF-IDF algorithm became popular because it prioritises terms
according to how often they show up in an article and how uncommon they are in other contexts.
Using evolutionary algorithms and ten text weighting features, Fattah and Ren (2008) were able to
accomplish summarization. It should be noted that study has not been confined to the English
language. While Budhi et al. (2015) used TF-IDF and virtual graphs to create an Indonesian text
summarizer, Aristoteles et al. (2013) used algorithmic genetics and 11 weighting features to build an
Indonesian summarizer. A sentence scoring system that utilised three approaches—word scoring,
graph scoring that analysed inter-sentence links, and scoring based on location and title similarity—
was developed by Ferreira et al. (2013), further advancing the discipline. Regardless of these

Page 3 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907

 Page 4 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907

advancements, there is still a challenge in selecting or combining strategies optimally for varied
settings, which is driving additional research in this intriguing topic.It summarises the original text
while keeping the flow of events and demonstrating the variety of approaches to automatic text
summarization..
C. Decision Tree
Automatic text summary determined by phrase frequency was first proposed by Luhn in 1958 and
later enhanced by Baxen-dale, who included sentence order as an additional factor. Edmunson built
from this in 1969 by incorporating additional features like title and keywords. Since then, the
discipline has seen the flourishing of diverse approaches. The reason the TF-IDF algorithm gained
traction is that it ranks terms by article frequency and their rarity in other contexts. To achieve
summarization, Fattah and Ren (2008) used evolutionary algorithms in conjunction with ten text
weighting characteristics. It is important to mention that research has not been limited to English. An
Indonesian textual summarizer was developed by Budhi et al. (2015) using TF-IDF and virtual graphs,
and by Aristotle et al. (2013) using algorithmic genetics and eleven weighting characteristics. To
further advance the field, Ferreira et al. (2013) created a phrase scoring system that used three
approaches: word netting, graph scoring (which evaluated inter-sentence linkages), and ranking
based on position and title similarity. Despite these developments, picking or combining techniques
optimally for different contexts remains a problem, which is motivating more research into this
fascinating subject.It displays the diversity of methods for automatic text summarization while
retaining the original text's sequence of events and providing a concise summary..

III. RELATED PAPER

A. A Look at ChatGPT's Restrictions on Summarising Text Based on Aspects or Queries [1]The
suggested method is an unsupervised structure for extractive text summarization. It integrates ranks
derived from several methods, such as topic, keyword, semantic, and positional data. The rank level
is more amenable to fusion than the score level owing to incompatibility and normalisation concerns,
hence it is where this fusion strategy chooses to take place.
B. Summarising Lengthy Documents Using Top-down and Bottom-up Approaches Using a single the
highest level with segments of fixed length, we infer [2] in this study. By combining token
representations, the segment representation are set up at the beginning.
Summarising Abstract Arabic Texts (C) Predicted on Deep Learning [3]Translating texts, analysing
emotions, and summarising texts are just a few examples of how deep learning approaches have
revolutionised vital tasks..
METHODOLOGY
The research itself uses four phases; the first phase is text documents collecting that will be used as
the data training and testing. The next is training phase to produce a model or rule for using decision
tree method. After it has a decision model, the next is testing phase to produce summarization
system. And the last phase is evaluation, which to test the accuracy level between summarization
system result and the manual summary.
Sentence scoring: Each sentence is assigned a score based on 8 features: TF/IDF (term
frequency/inverse document frequency), Uppercase words, Proper nouns, Cue phrases, Numerical
data, Sentence length, Sentence position, Similarity to title.
Decision tree: Sentence scores are used as training data for the decision tree algorithm. The tree is
constructed based on gain ratio, which measures the importance of attributes for classification. The
tree is used to select important sentences for the summary.
Evaluation: Precision, recall, and F-measure are used to evaluate the quality of the summaries.
Precision measures accuracy, while recall measures relevance. F-measure is a combination of
precision and recall
.
RESULTS

Page 4 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907

 Page 5 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907

Building a decision-model oriented text summarising system, the research utilises fifty written
materials containing news job titles, material, and manual descriptions (363 sentences). After that,
we use 8 criteria, such as TF/IDF, title similarity, and all capital letters, to give each for the 1237
sentences a weight. The following stage, which is to construct the decision tree model, requires this
data preparation. A decision tree has been constructed using the C4.5 algorithm, with features F1 as
the root, and is now prepared for rule generation. evaluation of the text summarising system using
fifty news articles revealed that the system's summaries were of comparable length to manually
overviews, having an average of eight sentences for the system and seven sentences for the manual,
and 168 words for the system and sixteen1 words for the manual, respectively. With an average F-
measure of 0.58 and a recall of 0.66, the method proved to be quite accurate. With a median the F-
value of 0.58 and a high score of 0.80, this research presents a phrase score and choice tree
approach to Indonesian text summarising. This method shows better accuracy for summarising
Indonesian text compared to prior approaches such as genetic algorithm (0.47 F-measure) and
residual Dirichlet allocation (0.55 F-measure)..
CONCLUSION
Summarising Indonesian text using a combination of sentence score and decision tree methods was
successfully demonstrated in this research. Eight features, including TF/IDF, sentence location, and all
capitalization, are used to determine a sentence's weight in sentence scoring. An important sentence
selection rule generator is an choice tree model that takes these weights and training data as inputs.
Although this study lays the groundwork for better Indonesian text summary, there are three
important areas where it could be improved even more. First, new, better rule models for identifying
key sentences might be created by building more complex scoring formulae for variables such TF/IDF
and sentence posi-tion. Second, multi-document summarising might be approached using a
promising structure of line scoring and decision trees. This approach would successfully summarise
not just each document but also collections of linked texts. Developing professional-quality
Indonesian text corpus and manual summaries would offer a valid standard for evaluating and
contrasting different text summarising algorithms. Lastly, a standardised testing and assessment
framework should be put in place. Automated Indonesian text summarization has great potential,
and by concentrating on these regions, research can build upon the success of previous work..
REFERENCES
[1] Xianjun Yang, Yan Li, Xinlu Zhang, Haifeng Chen, Wei Cheng, “Exploring the Limits of ChatGPT for
Query or Aspect-based Text Summarization,” University of California, Santa Barbara Microsoft.
[2] BoPang,ErikNijkamp,WojciechKrysci ́nski,SilvioSavarese,Yingbo Zhou, Caiming Xiong, Salesforce
Research.
[3] Y.M. Wazery, Marwa E. Saleh ,Abdullah Alharbi, and Abdelmgeid A. Ali,” Volume 2022 — Article ID
1566890.

Page 5 of 5 - AI Writing Submission Submission ID trn:oid:::1:2793903907