Hate Speech in Social Networks and Detection using Machine Learning Based Approaches

Abstract

The use of social networking sites has increased considerably in last few years and as a result the user generated contents in the web also increased manifold. These data are mostly present in unstructured and quasi-structured formats. Many social media platforms are being affected due to the presence of hate speech. It is present in many forms such as verbal aggression and through photos which we know as memes and so on. This study considers twitter data for detecting hate speech on the internet. From the past few years, machine learning and natural language processing approaches are being used to detect hateful content on the web. This study aims to deal with the problem of hate speech detection in text data using machine learning algorithms. Feature selection for the dataset was performed before passing the dataset to the machine learning models. Different machine learning algorithms are implemented on an open-source twitter dataset. Performance of different algorithms are compared using standard performance measure metrics and presented in this paper. The experimental result shows artificial neural network outperforms the other algorithms considered in this study.