You are on page 1of 35

Department of Computer Science and Engineering

VIGNAN’S LARA INSTITUTE OF TECHNOLOGY AND SCIENCE

Project Review
on
Analysis of Women Safety Using Machine Learning

Under The Esteemed Guidance of Presented By


Mr.D.SYAM KUMAR (M.tech) Batch No: B10
Assistant Professor K.CHANDRALEKHA 17FE1A0575
CSE Department P.N.V.DINESH 17FE1A0596
P.ANURADHA 17FE1A0599
P. KARTHIK 17FE1A0598

30/03/2021 Department of Computer Science &


1
Engineering
TITLE

Analysis of Women Safety Using


Machine Learning by Tweets

Department of Computer Science &


24/05/2021 2
30/03/2021 Department of Computer Science & Engineering
Engineering
2
CONTENTS

 Title
 Abstract
 Introduction
 Literature Survey on Papers
 Methodology

30/03/2021 Department of Computer Science & Engineering 3


ABSTRACT

• Women are experiencing lots of violence such as


harassment in places of several cities .
• We mainly focus on the role of social media which
can be used to promote the safety of women in India,
given the special reference to the participation of
many social media websites or applications such as
Twitter, Facebook and Instagram platforms .
• Also focuses on developing the responsibilities
among the common people so that the safety of
women around them is ensured.

30/03/2021 Department of Computer Science & Engineering 4


INTRODUCTION

• The main aim of this system is here we retrieve


comments, likes, shares and other forms of opinions of
people from multiple social networking sites.

• The system uses Sentimental Analysis for extracting


the sentiment behind any sentence or statement. It can
be called as a classification technique which is used to
obtain the opinion from comments.

30/03/2021 Department of Computer Science & Engineering 5


INTRODUCTION

• This paper also focuses on how a sense of


responsibility on part of Indian society can be developed
the common Indian people so that we should focus on
the safety of women surrounding them.

• Tweets on Twitter which usually contains images and


text and also written messages and quotes which focus
on the safety of women in Indian cities

30/03/2021 Department of Computer Science & Engineering 6


INTRODUCTION

• This can be used to read a message amongst the Indian


Youth Culture and educate people to take strict action
and punish those who harass the women.

• Other women share these messages and tweets which


further motivates other 5 men or 10 women to stand up
and raise a voice against people.

• Therefore, many researchers have used deep learning


and machine learning techniques to extract and mine
the polarity of the phrases.
Department of Computer Science &
01/20/2022 7
Engineering
30/03/2021 Department of Computer Science & Engineering 7
CONTENTS

 Literature Survey Paper-1


 Literature Survey Paper-2
 Literature Survey Paper-3
 Literature Survey Paper-4

Department of Computer Science &


24/05/2021 8
15/04/2021 Department of Computer Science & Engineering
Engineering 8
LITERATURE SURVEY PAPER -1

• (Pang and Lee, 2004)proposed an approach to


automatically detect sentiments on Twitter messages
(tweets) that explores some characteristics of how
tweets are written and meta-information of the words
that compose these messages.
• Features like unigrams, bigrams, part-of-speech tags
are used.The main limitation of this approach is the
cases of sentences that contain antagonistic
sentiments. 

01/20/2022 9
15/04/2021 Department of Computer Science & Engineering 9
LITERATURE SURVEY PAPER -1

Robust Sentiment Detection on Twitter from


.  Biased and Noisy Data

• We implement a 2-step
sentiment detection framework.
• Subjectivity detection
• Polarity detection

Department of Computer Science &


24/05/2021 10
15/04/2021 Department of Computer
EngineeringScience & Engineering 10
LITERATURE SURVEY PAPER -1

BLOCK DIAGRAM

15/04/2021 Department of Computer Science & Engineering 11


LITERATURE SURVEY PAPER -1

ADVANTAGES:

It is k appa coefficient.
 It is Effective and robust sentiment detection approach.

DISADVANTAGES:

 It Antagonistic sentiments.
Rich literature in sentiment detection.

Department of Computer Science &


01/20/2022
24/05/2021 12
15/04/2021 Department of Computer
EngineeringScience & Engineering 12
LITERATURE SURVEY PAPER -2

 Twitter and Instagram broadly spread the messages


across the whole globe and cause the ladies to don't
hesitate to communicate their perspectives and
sentiments.
  Utilizing this we can know the condition of their
psyche when they go out for work or travel in a
public vehicle or encompassed by mysterious men.AI
calculation assists with getting sorted out and
perform investigation..

24/05/2021 13
15/04/2021 Department of Computer Science & Engineering 13
LITERATURE SURVEY PAPER -2

BLOCK DIAGRAM

Department of Computer Science &


24/05/2021 10
15/04/2021 Engineering
Department of Computer Science & Engineering 14
LITERATURE SURVEY PAPER -2

ADVANTAGES:

.  Twitter has arisen for a blog and person to person communication with
manages known as “tweets” and also can share our views& feelings in
this platform.
 By these tweets we can find out about the situation with ladies well
being in Indian urban communities .

DISADVANTAGES:

 There are difficulties in understanding the feeling in the sentences.


 The way toward eliminating retweets and repetitive information as it
was not having a reasonable and unique perspective on security status.

Department of Computer Science &


01/20/2022
24/05/2021 15
15/04/2021 Department of Computer
EngineeringScience & Engineering 15
LITERATURE SURVEY PAPER -3

  Like Yu and Hatzivassiloglou (2003)proposed


automatically finding the people who hold opinions
about that topic and the sentiment of each opinion.

The topic is divided into sentences and using pos


tagger the sentences are divided into noun , adjective
and verbs , the problem is Even in a single sentence, a
holder might express two different opinions.
this system only detects the closest one.

15/04/2021 Department of Computer Science & Engineering 16


LITERATURE SURVEY PAPER -3

BLOCK DIAGRAM OF
Determining The Sentiments Of Opinions

Department of Computer Science &


24/05/2021 17
15/04/2021 Department of Computer Science &
Engineering Engineering 17
LITERATURE SURVEY PAPER -3

ADVANTAGES:
• Supervised Learning is so accurate.
• By Unsupervised Learning work done is less.
• Hybrid Approach achieve better accuracy.

DISADVANTAGES:
• Unsupervised Learning is not so accurate when compared so
supervised learning.
• In Supervised Learning sarcastic expressions are hard to understand.
• Emotions are also hard in analyzing especially in supervised
learning.

Department of Computer Science &


01/20/2022
24/05/2021 18
15/04/2021 Department of Computer Science &
Engineering Engineering 18
LITERATURE SURVEY PAPER -4

• Twitter sentiment analysis is an application of sentiment


analysis on data from Twitter(tweets), in order to
extract sentiments conveyed by the user. 

• In this they used Nltk (Natural Language Processing) is


a library used for Text processing and classification .

• Scikit also a library used for classification and Numpy


is a package , The problem in this project is it still lacks
the dimension of diversity in the data
Department of Computer Science &
01/20/2022 19
Engineering
15/04/2021 Department of Computer Science & Engineering 19
LITERATURE SURVEY PAPER -4

Twitter Sentiment Analysis using Machine Learning Algorithms


on Python 

Department of Computer Science &


24/05/2021 20
15/04/2021 Department of Computer Science &
Engineering Engineering 20
LITERATURE SURVEY PAPER -4

ADVANTAGES:
 By Twitter sentiment analysis in python results in a high accuracy.

 By these tweets analysis they reach the efficiency of almost 85-


90%.

DISADVANTAGES:
 But it still lack the dimension of diversity in the data. Along with it
has a lot of application issues with slang used and short forms of
words.
 Also it still don’t perform well when the number of classes are
increased.
Department of Computer Science &
01/20/2022
24/05/2021 21
15/04/2021 Department of Computer Science & Engineering
Engineering 21
EXISTING METHODS
• In the recent years a large number of people have been attracted towards

social media platforms like Face book, Twitter and Instagram point and most

of the people are using it to express their emotions and also their opinions .
• There are several method of sentiment that can be categorized like machine

learning hybrid and lexicon-based learning. Also there are another

categorization janta presented with categories of statistical, knowledge-based

and age wise differentiation approaches


• The accuracy of the Twitter analysis and prediction can be obtained by the

use of behavioral analysis on the basis of social networks.

01/06/2021 Department of Computer Science & Engineering 22


EXISTING METHODS

01/20/2022 23
01/06/2021 Department of Computer Science & Engineering 23
PROPOSED METHOD
• As People communicate and share their opinion actively on
social medias including Face book and Twitter, Social network
can be considered as a perfect platform to learn about people’s
opinion and sentiments regarding different events.

• Since Twitter contains short texts, people tend to use different


words and abbreviations. These phrases are difficult to extract
their sentiment by current NLP systems easily.

• Therefore, many researchers have used deep learning and


machine learning techniques to extract and mine the polarity of
the phrases

Department of Computer Science &


01/20/2022 24
01/06/2021 Department of Computer Science & Engineering
Engineering
24
PROPOSED METHOD

Our proposed system consists of

• Random Forest Algorithm


• Support Vector Machine Algorithm.

Department of Computer Science &


01/20/2022 25
01/06/2021 Department of Computer Science & Engineering
Engineering
25
METHODOLOGY

Department of Computer Science &


01/20/2022 26
01/06/2021
Engineering
Department of Computer Science & Engineering 26
METHODOLOGY
RANDOM FOREST:
 Random Forest is an ensemble learning algorithm for
classification and regression. Random Forest generates a
multitude of decision trees classifies based on the aggregated
decision of those trees.

 Random Forest Classifier from sklearn . ensemble provided


by scikit-learn. We experimented using 10 estimators (trees)
using both presence and frequency features. presence
features performed better than frequency though the
improvement was not substantial

Department of Computer Science &


01/20/2022 27
01/06/2021
Engineering
Department of Computer Science & Engineering 27
RANDOM FOREST WORKING:
Step-1: Select random K data points from the
training set.
Step-2: Build the decision trees associated with
the selected data points (Subsets).
Step-3: Choose the number N for decision trees
that you want to build.
Step-4: Repeat Step 1 & 2.
Step-5: For new data points, find the predictions
of each decision tree, and assign the new data
points to the category that wins the majority votes.
Department of Computer Science &
24/05/2021 28
Engineering
01/06/2021 Department of Computer Science & Engineering 28
METHODOLOGY

Department of Computer Science &


01/20/2022 29
01/06/2021
Engineering
Department of Computer Science & Engineering 29
SUPPORT VECTOR MACHINE:
A support vector machine (SVM) is a
supervised machine learning model that
uses classification algorithms for two-group
classification problems. After giving an SVM
model sets of labeled training data for each
category, they’re able to categorize new
text.

Department of Computer Science &


01/20/2022
24/05/2021 30
01/06/2021 Department of Computer Science & Engineering
Engineering 30
SUPPORT VECTOR MACHINE WORKS:

Department of Computer Science &


24/05/2021 31
01/06/2021 Department of Computer Science &
Engineering Engineering 31
Step-1:  Support vector machine creates a new classifier.
Step-2:  Select how you want to classify your data ( specifying
positives and negative words).
Step-3:  Import your training data(twitter data).
Step-4:  Define the tags for your SVM classifier(either positive or
negative or neutral).
Step-5:  Tag data to train your classifier.
Step-6: Test your classifier and integrate topic classifier.
Step-7: Fisher kernel(svm classifier) classifier classifies the data
based on training data.
Step-8: Displays the result after evaluation.

Department of Computer Science &


24/05/2021 32
01/06/2021 Department of Computer Science &
Engineering Engineering 32
APPLICATION

• Instagram
• Facebook
• Other social media

01/06/2021 Department of Computer Science & Engineering 33


Any Queries??

24/05/2021 computer science engineering 34


01/06/2021 Department of Computer Science & Engineering 34
Thank You

01/06/2021 Department of Computer Science & Engineering 35

You might also like