Aggression Identification

Aggression Identification is a crucial task in the field of natural language processing, aiming to develop models that can classify text data into three categories: implicit aggression, explicit aggression, and non-aggressive text. This task leverages the TRAC-2 dataset, which includes 5,000 annotated data points from social media, covering Bengali (in Roman and Bengali script), Hindi (in Roman and Devanagari script), and English, to enhance the model's multilingual recognition capabilities, making it socially valuable.