Cyberbullying and toxic language detection on social media for Bangla language

Abstract

As more people use social media, toxic language and cyberbullying become more common with the Bengali-speaking community particularly. The complexity of Bangla text data makes it difficult for traditional natural language processing (NLP) algorithms to identify harmful content. This study proposes a machine learningbased solution that recognizes and categorizes harmful language and “Cyberbullying in Bangla text on social media”, leveraging BanglaBERT’s advanced features. As more people use social media, toxic language and cyberbullying are on the rise, with the Bengali-speaking minority particularly vulnerable. The proposed machine learning-based solution achieved 94% testing accuracy in detecting and categorizing cyberbullying and offensive language on digital platforms that support Bengali.