IRJET - Sentiment Analysis of Posts and Comments of OSN
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 03 | Mar 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 5364
Sentiment Analysis of Posts and Comments of OSN
Abhishek Chaube1, Vaidehi Dani2, Trupti Dhapola3, Prof. Madhura Vyawahare4
1,2,3Student, Dept. of Information technology Engineering, Pillai College of Engineering, Maharashtra, India
4Professor, Dept. of Computer Engineering, Pillai College of Engineering, Maharashtra, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - Online social networking sites play an important
role in the day to day life of everyone. Twitter is one of the
most used platforms for social interaction. Twitter allows its
users to have their own views, comments, and emotions
majorly using texts and emoticons. There are many views,
posts, and comments which may hurt the sentiments of
other people. Monitoring such information over the
discussion forums or groups is required but it becomes
difficult as each individual has their own opinions and
suggestions. Nowadays, depression-related posts have
become a major concern. The proposed system aims to
identify the sentiment behind the content posted on twitter
to find depressed people. With sentiment analysis, we are
targeting to find out the user’s positive emotions (like
happy, excited, surprised, satisfaction) and negative
emotions (like depression, anxiety, stress, sad, angry). Then
negative sentiments will further, be classified and the degree
of depression is displayed. This study is significant to
monitor the emotions of the users.
Keywords- Sentiment analysis, emotions, degree,
comments.
1. INTRODUCTION
The project is based on Machine Learning and Data mining
domain. Machine Learning is a field of computer science
that gives the computer the ability to learn without being
explicitly programmed. Machine Learnings’ main focus is
to provide algorithms that can be trained to perform a
task. Data mining is the process of discovering patterns in
large data sets involving methods at the intersection of
machine learning, statistics, and database systems. Data
mining is an interdisciplinary subfield of computer science
and statistics with an overall goal to extract information
(with intelligent methods) from a data set and transform
the information into a comprehensible structure for
further use.
2. LITERATURE SURVEY
Sentiment Analysis
The author Neethu and Rajashree explained the need for
sentiment analysis to match social media usage. Social
media includes all the sites such as blogs, Facebook,
Twitter, and other social networking sites. Sentiment
score helps examine whether the tweet seems to be
positive, Negative or Neutral. They have gathered reviews
on services and brands is done using opinion mining. The
marketing-related sector is still at the forefront followed
by finance, hospital, and tourism. Politics and Government
are still a few of the bodies that are beginning to use
sentiment analysis [1].
Sentiment analysis is done to detect the level of
depression in paper [2]. The keyword-based approach is
used to calculate the intensity of emotions. Keyword-
based approach classifies text based on the presence of
negative or positive polarity words such as happy, joyful,
delighted, miserable, sad, terrified, and uninterested,
happy, joyful, respectively. Weights are given to different
words related to depression by evaluating the polarity of
those words and the result is displayed in the metric form.
The author explains that Social Media has become very
popular as a way of communication. In this research,
authors have used a method for Automatic collection of
data that can be used to train sentiment classifiers. They
observe different emotions such as positive, negative,
neutral, etc. They used different methods such as Tree
Tagger, Syntactic Structures, N-gram and POS-tags
amongst which POS-tags are able to identify sentiments
strongly from emotional text [3].
The authors have built a model that analyzes sentiments
on Twitter using Machine Learning techniques. They have
applied Bigram, Unigram, Object-oriented features as an
effective feature set for sentiment analysis. They have
used a data set about 200000 tweets for training
classifiers. They built a sentiment analysis model based on
supervised learning such as Naive Bayes and Support
Vector Machine for enhancing elective classifIcation.
Twitter APIs are used as a library tool to collect tweets
from the internet for sentiment analysis and a system is
built based on Naive Bayes (NB) and Support Vector
Machine (SVM). Accuracy is not 100% as there are many
grammar mistakes and repetitive words that disturb the
feature dataset. This paper only works for English words
and not any other languages [4].
Twitter sentiments are considered and analyzed using
different Symbolic and Machine Learning techniques to
identify sentiments from text. There are certain issues
while dealing with identifying emotional keywords from
tweets having multiple keywords. It is also difficult to
handle misspellings and slang words. To deal with these
issues, and efficient feature vector is created by doing
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 03 | Mar 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 5365
feature extraction in two steps after proper preprocessing.
In the first step, twitter specific features are extracted and
added to the feature vector. After that, these features are
removed from tweets and again feature extraction is done
as if it is done on the normal text. These features are also
added to the feature vector. Classification accuracy of the
feature vector is tested using different classifiers like
Naive Bayes, SVM, Maximum Entropy and Ensemble
classifiers [5].
How depression can be monitored by users using social
media is explained in paper [6]. Sentiment analysis and
affective computing methods are used to detect and
monitor depression. This uses supervised and
unsupervised learning methods and a lexicon-based
approach is used. It detects depression and also provides a
multimodal system to check facial expressions, images,
and videos shared by the users [6].
Paper [7] aims to identify the opinion mining and
sentiment analysis components for extracting both English
and Malay words on Facebook. Information, in terms of
texts, are extracted and clustered into emotions. This work
begins with transforming unstructured information into
meaningful lexicons after extracting Facebook's contents.
All of the meaningful lexicons are stored in a database
after manual identifications are carried out. With
sentiment analysis, emotions are classified into happy
(positive), unhappy (negative) and emotionless [7].
User activities and interactions in the tourism domain are
analyzed. In particular, the emotions of the users
regarding their forthcoming trips are studied with the
objective to characterize interdependencies between
them. Social network analysis is applied to examine
interactions between the users. To capture their emotions,
text mining techniques, and sentiment analysis is applied
to construct a measure, which is based on free-text
comments in a travel forum [8].
3. PROPOSED SYSTEM
The system architecture is given in Figure 1. Each block is
described in this Section.
Fig- 3.1: Proposed System Architecture
Data collector: Personalization is a concern about
adapting to interests, comments, emoticons, and
preferences of users on social networking sites to collect
data scrapped by such websites.
Noise filter: The filtered out sentences can be further
narrowed down to only some specific words that can be
triggering or words that are required to further carry out
testing.
Smart filter: The words that are filtered out are then
compared with words that are stored and that commend
different emotions.
Result viewer: The data that was mined from different
social networking sites after being filtered, compared and
then sorted according to a specific emotions is then
measured down and calculated according to every
emotion which is then calculated all together to get a score
which will then specify the percentage of each emotion
and the ranking of the same.
4. IMPLEMENTATION
4.1 Software
Operating
System
Windows XP Professional With
Service pack 2
Programming
Language
Python
Database Oracle 9
Table- 4.1: Software details
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 03 | Mar 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 5366
4.2 Hardware
Processor 2 GHz Intel
HDD 180 GB
RAM 2 GB
Table- 4.2: Hardware details
4.3 Inputs Details
Standard datasets are taken from Tweepy. Tweepy is used
to get live data from twitter. The input data will be tweets
extracted from the user. Firstly it takes the consumer key,
consumer secret, access key and access secret from twitter
developer available easy for each user. These keys will
help the API for authentication. Then the user is asked to
enter the keyword and number of tweets to search about.
Fig -4.1: Input data
4.4 Output Details
In order to evaluate the proposed system, experiments
were conducted on tweets related to the keyword entered
by the user. The data was collected from hundreds of users
who use twitter to express their views and opinions.
Firstly, it shows the authentication of the twitter
developer account. It gives the date and time and other
basic information about when the account was created.
Then it displays the tweets based on the search keyword.
It will show the number of tweets the user wants to see.
Fig -4.2: Output Data
5. CONCLUSIONS
The application helps up find sentiments related to a
specific subject, argument, problem, the issue using
keywords associated to topic of interest. The data that is
the sentiment of social media users is collected and
analyzed. This data can be used for various purposes such
as consumer requirement analysis, impact caused on users
due to certain issues and so on. The analyzed data then
can be used to improve user experiences in the future.
6. ACKNOWLEDGEMENT
We would first like to thank our Project Coordinator and
our guide, Prof. Madhura Vyawahare, who keeps
encouraging us to do such projects which helps improve
practical knowledge. We also thank her for solving our
queries and helped us through this project.
We also thank our project coordinator, Prof. Gayatri
Hegde for her guidance on the planning and
implementation of our project.
We give special gratitude to our Principal, Dr. Sandeep
Joshi, who always encouraged and motivated us to do
innovative things that will improve our knowledge.
We also thank our H.O.D of Information Technology
Department, Dr. Satishkumar Varma, for this
opportunity of practically implementing the concepts we
study, through Project. This would not have been possible
without the opportunity. We are thankful to all who
provided us an opportunity to complete this project.
REFERENCES
[1] Neethu M S, Rajasree R,” Sentiment Analysis in Twitter
using Machine Learning Techniques”(2013)
- 4. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 03 | Mar 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 5367
[2] Shahid Shayya, Noor Ismavati Jaffar, "Sentiment
Analysis of Big Data: Methods, Applications, and Open
Challenges"(2016)
[3] Young sub-Han,” Sentiment Analysis on Social Media
Using Morphological Sentences Pattern Model”,(2017)
[4] Alexander Pak, Patrick Paroubek, "Twitter as a Corpus
for Sentiment Analysis and Opinion Mining"(2017)
[5] Neethu MS And Rajasree R, "Sentiment Analysis in
Twitter using Machine Learning Techniques"(2016)
[6] Chiara Zucco, Barbara Calabrese, Mario Cannataro,
"Sentiment Analysis and Affective computing for
depression monitoring." (2017)
[7] N. Azmina, Nasiroh Omar, "Sentiment Analysis:
Determining people’s emotions in Facebook" (2017)
[8] Julia Neidhardt, Hannes Werthner," Predicting
happiness: user interactions and sentiment analysis in an
online travel form"(2018)
[9] S. Alami, O. Elbeqqali “Detecting Suspicious Profiles
Using Text Analysis Within Social Media” Journal Of
Theoretical And Applied Information Technology,(2017)