Hate Speech Detection for Micro-blogging Websites

Abstract

Hate speech detection is a problem of filtering negative textual content on social media. I have limited this problem to content on micro-blogging website Twitter.Input to our problem is collection of tweets by various users. Before applying any classification or clustering approach to solve the problem data needs to be in specific form.For that Regex library of python is used. Another python library NLTK is used. As the data is in text format we have use feature extraction techniques which gave us set of features which can be fed to classification algorithm. The classification algorithm that is used for this problem in this report is Naive Bayes Classification, Support Vector Machine and Random forest algorithm.Accuracy measurement for all the algorithm is done simply in the form of Precision and Recall. Other learning approach like LSTM network,deep learning etc can be used which tend to give higher accuracy than simple classification approach.This leads to Future scope for improving this problem.