Performance Analysis of Supervised Machine Learning Algorithms for Detection of Cyberbullying in Twitter

Abstract

AbstractThese days, the use of social media is inevitable. Social media is beneficial in several means, but there are severe terrible influences. A crucial difficulty that needs to be addressed is cyberbullying. Social media, especially Twitter, advances numerous concerns due to a misunderstanding concerning the notion of freedom of speaking. One of those problems is cyberbullying, which influences both man or woman victims as well as the societies. Harassment by way of cyberbullies is a big issue on social media. Cyberbullying affects both in terms of the mental and expressive manner of someone. So there is a need to plan a technique to locate and inhibit cyberbullying in social networks. To conquer this condition of cyberbullying, numerous methods have been devised. This paper would help to comprehend the methods and procedures like logistic regression (LR ), naïve Bayes (NB ), support vector machine (SVM), and term frequency—inverse document frequency (TF-IDF) which are used by numerous social media web sites, especially Twitter. In this paper, we have worked on the accuracy of the SVM, LR, and NB algorithms to detect cyberbullying. We observed that the SVM outperforms the others.KeywordsCyberbullyingSocial networksMachine learningTwitterVictimsLogistic regression (LR)Naïve Bayes (NB)Support vector machine (SVM)