Promoting Constructive Online Debates Through Toxic Comment Classifier

The widespread use of online platforms for communication has given rise to the issue of toxic comments, which can have detrimental effects on individuals and communities. In this research, we present a toxic comment classifier that aims to automatically identify and classify toxic comments with high accuracy. We leverage a dataset consisting of annotated toxic comments, employing a mix of machine learning methods and natural language processing techniques. Through comprehensive testing and evaluation, we show that our method is effective in correctly identifying harmful comments. Our classifier achieves F1-score exceeding 85% across different toxicity categories. The developed toxic comment classifier has the potential to contribute to the creation of safer and more inclusive online environments, enabling proactive moderation and targeted intervention. This study emphasises how crucial it is to use machine learning approaches to fight toxic behaviour and promote constructive online discussion.