International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 08 | Aug 2023
p-ISSN: 2395-0072
www.irjet.net
Automatic Prediction and Countering of Communal Tweets using Machine Learning Techniques R. Durga1, V. Pavithra2 1Student, M.E, Dept. of Computer Science and Engineering, T.J.S Engineering College, Tamil Nadu, India
2Assistant Professor, Dept. of Computer Science and Engineering, T.J.S Engineering College, Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Social networks such as Twitter and Facebook
calamities is focused. While looking through these tweets, it has been observed that a large amount of communal tweets are posted rather than anti communal tweets. A calamity generally affects the morale of the masses making them vulnerable. Often, taking advantage of such situations, hatred and misinformation are propagated in the affected region, which may result in serious deterioration of law and order.
are seriously affected by abusive and offensive content. A lot of research has been carried out in recent years for automatic identification of different types of hate speech. Hate speech can come under several categories such as gender, ethnicity, nationality, racial groups and religions of the targeted groups. This work explores methods which aim to counter potentially harmful communal tweets which are pointed towards religious groups such as Hindus, Sikhs, Muslims, Jains, Christians, Buddhists. Communal tweets are largely posted during calamities and taking advantage of such situations, hatred and misinformation are propagated in the affected region which may result in serious riot and violence. Considering the potentially adverse effects of communal tweets, a classifier to identify communal tweets and non communal tweets is developed using the Support Vector Machine algorithm which performs better than existing approaches. Users who post communal tweets are identified. Anti communal tweets are identified from the non communal tweets using a classifier and by finding the appropriate anti communal tweet to a communal tweet, this system counters communal tweets by making use of such anti communal tweets posted by some users. The proposed system will be really helpful to reduce the increasing communal venom in society by blocking the users who repeat posting such communal tweets. It also changes the negative opinion of communal tweeters by promoting anti communal contents.
A detailed analysis of communal tweets during disaster situations are detected and countered. This system characterizes users who initiate and promote communal tweets and also suggest a way to counter such communal content with anti communal tweets that asks users not to spread communal hatred. Considering the adverse effects of communal tweets, a Support Vector Machine Classifier is developed to automatically separate the communal tweets from the non communal ones. After identifying communal tweets, the nature of communal tweets and the users who post them are identified. Initiators, who initiate the communal tweets are identified and propagators who retweet the communal tweets posted by initiators or copy the contents of other initiators and post their own tweet with minor changes are also identified. Apart from the communal tweets, non communal tweets are analyzed to automatically identify anti communal tweets for countering. This will dissuade people from posting communal contents. Hence, a convincing way to counter the communal tweetis is to counter them with anti communal tweets to the initiators and propagators who post communal tweets. A real time system is also developed that automatically collects tweets from the user and counters with anti communal tweets if the posted tweet is communal and blocks the users who repeat posting the communal tweets through which communal venom in the society can be reduced.
Key Words: Anti communal tweets, Classifier, Communal tweets, Religious groups, Users
1. INTRODUCTION In recent years, there is an increase in propagation of communal speech on social media and the urgent need for effective countermeasures are needed. Twitter is a social broadcast network that enables people to publicly share brief messages instantly around the world. This message brings a variety of people with different voices, ideas and perspectives. Twitter prevents and filters only abusive contents and offensive words from the tweets but not the communal contents. Using this, communal tweets are pointed towards certain religions, racial communities, politicians or certain groups. Communal tweets are posted during times of calamities or emergency situations. Hence, analyzing the communal tweets on twitter during
© 2023, IRJET
|
Impact Factor value: 8.226
2. EXISTING SYSTEM It has been observed that offensive tweets are often posted during calamities to which the attackers are affiliated. However, it is quite surprising that in certain geographical regions such as the Indian subcontinent, communal tweets are posted even during natural disasters such as floods and earthquakes and also during man-made disasters. Several studies have attempted to identify online content that is potentially hate speech or offensive
|
ISO 9001:2008 Certified Journal
|
Page 758