Troll Detection and Anti-Trolling Solution using Artificial

Abstract

With the increase in usage of social media platforms, due to which trolling and use of abusive language has burgeoned proportionately. The sole reason for this is that there is no surveilling authority on these platforms. Anyone from kids, teenagers to adults can fall prey to trolling. This paper focuses on using Artificial Intelligence and Machine learning algorithms to invigilate such bullies and further classify them for enhanced analysis. We will be introducing lexical, aggression, syntactic and sentiment analyzers to examine the data and determine if it was meant to be a troll or not. The output of these analyzers will be then fed to algorithms such as Naive Bayes and classifiers like Decision Tree, Random forest, Multinomial, Logistic regression to segregate the trolls in different categories like offensive, targeted, individual, group etc and use visual representation tools to improve the analysis.