Comparative Analysis on Joint Modeling of Emotion and Abuse Detection in Bangla Language

Abstract

AbstractEmotions are not linguistic entities, although they are easily articulated through language. Emotions influence our actions, ideas, and, of course, how we communicate. On the other hand, abusive text, such as undiscriminating slang, offensive language, and vulgarity, is more than just a message; it is a tool for very serious and brutal cyber violence. Hence, detection of such language has become very important in any language now-a-days. Therefore, many works and researches have been done on detecting emotional language, abusive language or both in many dialects including Bangla. This paper proposes to present a comparative analysis of different researches made on detecting emotional and abusive Bangla language. It further aims to present the best approach that tailors certain attributes of emotional and abusive language detection with respect to their prognosis performance and their implementation toughness in Bangla lingo. Potential enhancements for future study are presented in the paper, while the limitations of current researches are addressed and discussed. This work seeks to bring a fresh viewpoint to the joint modeling of emotional and abusive language detection in Bangla by examining and criticizing flaws and in order to offer future changes, poor design choices must be examined.KeywordsNLPCNNRNNMTLBiGRULSTMICTAI