A Framework for Actor-Oriented Automated Hate Speech Detection

Abstract

The development of automated hate speech detection research has not yet detailed the actor who is the source of hatred, even though the forms of hate speech intervention for perpetrators and supporters are different. Previous research on this topic has paid much attention to the action component in the form of sentimental words. However, it has yet to pay attention to the actor component that can differentiate legal hate speech from illegal hate speech and transforms hateful or offensive speech into free speech, thus dealing with misclassification, as reported in a previous study. This research proposes a framework for the automated detection of hate speech that provides an actor and action component to solve the problem of such errors and meets the need for comprehensive interventions. This study shows how to apply the framework in a rule-based approach by considering the actor component in predicting hate speech and differentiating it from others. The prediction process is called actor-oriented automated hate speech detection.