查看文章

Social network usage is growing daily, making it impossible to manage the enormous amount of data being generated. The presence of abusive behavior and hate speech is a clearly harmful phenomenon that is evident on these networks. Due to its importance, recent studies have revealed a significant concern in this field. This review aims to provide insight into the tasks and procedures associated with the automatic detection and classification of texts containing hate speech. As the domain of hate speech is wide, an analysis of definitions is conducted and a comprehensive and unifying definition is proposed. This paper also investigates the latest datasets across languages used to train AI models. Recent studies show that feature selection plays a key role in detecting hate speech. In this research, we analyze which are the most utilized and impactful features from works done in this domain. While various …