Abstract

Social media provides a user-friendly platform for interested persons or groups to express opinions and discuss freely their topics of interest which enhances the propagation of online hate speech, which is considered a serious issue in the web community because cyber hate speech has the potential to cause harm to individuals and society at large. The main objective of this paper is to study current literatures on the detection of online hate speech to determine the trends in online hate speech detection tasks. Various databases (Elsevier, IEEE Xplore, ACM digital library, Springer, and Google Scholar) were searched to obtain the materials used for this review. The method adopted for this review is the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Various databases (ScienceDirect, IEEE Xplore, ACM digital library, Springer, and Google Scholar) were searched. A total of 31,714 publications from 2015 to 2020 were studied, a total of 31,673 papers were excluded based on exclusion criteria, and 41 papers were included based on inclusion criteria. The results show that the Support Vector Machine learning algorithm was the most commonly used algorithm for online hateful text classification, though, deep learning algorithms and hybrid deep learning approaches are gaining grounds recently. This paper concludes that machine learning and deep learning approaches have proven effective in the classification of hateful text on social media. However, there is a need for the development of hybrid cross-platform models for hate speech detection and blocking.