International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 05 | May 2022
p-ISSN: 2395-0072
www.irjet.net
Election Result Prediction using Twitter Analysis Ajay Rao1, Varun Kanade2, Chinmay Motarwar3, Prof. Shital Girme4 1, 2, 3Undergraduate
Student, Dept. Computer Engineering, Pune Institute of Computer Technology, Pune, India Dept. Computer Engineering, Pune Institute of Computer Technology, Pune, India ---------------------------------------------------------------------***--------------------------------------------------------------------4Professor,
Abstract - Elections in India has always been considered
Elections play an important role in a democratic country. Indian parliamentary system gives its people the right to decide who will govern them for the next five years. During the tenure of Feb 22 to March 22, five state elections are lined up, with the important one being at Uttar Pradesh, which sends the largest number of MPs to parliament. The major national political parties contesting in the elections are Bhartiya Janata Party(BJP), Indian National Congress (INC), Aam Aadmi Party(AAP), Samajwadi Party(SP), Shiromani Akali Dal(SAD) and Naga People’s Front(NPF).
as an important event and has been keenly followed by majority of people. The rapid increase of social media in the recent past has provided end users a powerful platform to voice their opinions. Twitter, being one such platform, provides day-to-day updates on political events through different hashtags and trends. People provide their opinion by reacting on such political events. Our approach is to gather a collection of tweets of top political parties contesting within the General State election, 2022, then compute the sentiment score. Dataset contains mixture of both popular as well as recent tweets related to specific political party. Specific keywords are used to extract tweets for a party like ‘BJP elections 2022’, ‘#UPelections BJP’, ‘#Punjabelections BJP’, etc. We utilized a combination of VADER Sentiment Analyzer and classic machine learning algorithms like Random Forest Classifier, SVM, etc. to build our classifier and classify the test data as positive, negative and neutral tweets. Therefore, this work analyses tweets collected from twitter and predicts election results by performing sentimental analysis on them.
2. LITERATURE SURVEY Parul S and Teng-Sheng Moh [2] predicted the results of 2016 Indian general elections using tweets in Hindi language. They performed text mining on 42,235 tweets collected over a month. They applied three ML algorithms. The accuracy of the Naïve Bayes’ algorithm was 62.1% and the accuracy of Support. Vector Machine was 78.4%. Final prediction was done by utilizing SVM, since the accuracy was higher.
Key Words: Sentimental Analysis, Twitter, Supervised learning, Natural Language Processing, Machine Learning
Dr D. Rajeswara Rao and team [3] gathered a dataset of more than 500,000 tweets out of which 80% was used for training and rest for testing. They predicted which political party had more influence on social media. Proposed a system that trained the dataset for more than 2 days and a classifier was built. Experiments proved that SVM was the most accurate model built with an accuracy of 80%.
1. INTRODUCTION Social media has become a powerful tool for sharing opinions. There are social media platforms like Facebook, Twitter and Google+ to share opinions, reviews and ratings. All major political parties and their members all over the world have their official accounts on Twitter with millions of followers. They consider this platform as a medium to connect with young people who might vote them. With significant rise of Indian users on Twitter during the pandemic, people have been more vocal to criticize or appreciate a political decision.
Ferdin Joe and John Joseph [4] used decision tree to predict 2019 Indian General Elections. The results obtained in the proposed methodology showed it had a promising future in predicting Indian election results. Meng-Hsiu Tsai and his team [5] at Middle Georgia State University presented a machine learning strategy to analyse Twitter data for predicting the results of local elections in US. They categorized their results into 5 classes namely very positive, positive, neutral, negative and very negative. They used RNTN model to calculate weighted sentiment scores.
Sentimental Analysis is a method to teach a machine to extract emotion from a given text [1]. A text can be anything, a simple review, a social statement, tweets or messages. Twitter Sentiment Analysis of tweets regarding elections can be used by the general public as well as the political parties to understand the positive or negative views of people regarding a particular political party, thus, helping to predict the election results during that period.
© 2022, IRJET
|
Impact Factor value: 7.529
Payal Khurana Batra and her team [6] predicted election results of Lok-Sabha 2019. After pre-processing, they split the data into two parts containing BJP and Congress tweets in separate sets. They trained their model using
|
ISO 9001:2008 Certified Journal
|
Page 2357