Cyberbullying Sentence Detection Using Machine Learning for the Bengali Language

Abstract

Cyberbullying has become a significant issue in recent years, particularly among young people, and there is a need for practical tools to detect and prevent it. Cyberbullying is sending offensive, abusive or threatening messages to insult a person. It is more dangerous than traditional bullying because it can occur at any time and from any location and be done anonymously. Social media is getting vast amounts of data every day. We see in current trends that Cyberbullying is a bigger and bigger problem. It is even more severe than regular bullying’s on the internet, like Facebook, Twitter, or other internet platforms. Finding a bully is far more challenging. So, we collect a data tool that essentially interacts and communicates with various social media site data using vest technologies like natural language processing and machine learning that automatically detect bullying. The study employs various machine-learning algorithms to develop a model for detecting cyberbullying sentences in Bengali text also discusses the challenges faced in developing a machine-learning model for Bengali cyberbullying detection and the potential solutions. Overall, the study demonstrates the potential of machine learning for detecting cyberbullying in Bengali and contributes to developing practical tools to prevent and combat cyberbullying in the Bengali-speaking community. This paper proposes an approach to detect cyberbullying in Bangla sentences using social media datasets and machine learning techniques, and the evaluation dataset shows that Multinomial Naïve Bayes performs better and achieves an accuracy of 79.49%.