A short review on hate speech detection: challenges towards datasets and techniques

Hate speech detection can be seen as the process of categorizing and identifying speech or written content that promotes discrimination, or violence towards individuals or groups. It can include various computational algorithms, natural language processing methods, and machine learning algorithms to automatically analyze and classify text data for the presence of hate speech. In this paper, we present the review of several approaches for hate speech detection and their challenges towards datasets and techniques. We highlight the examples of the architectures for hate speech detection based on combination of classical and deep learning classifiers. Moreover, we illustrate the examples of multi-labels for hate speech detection from various datasets. We found that there are still issues with defining and detection hate speech. Thus, the investigation of machine learning methods for hate speech detection is still a fruitful research area.