Finding Harmful Comments on Social Networking Sites Using NLP and Machine Learning Methods

Abstract

The usage of violent language has significantly increased due to social media and networking. A key component in this is the younger generation. More than half of young people who use social media are affected by cyberbullying. Harmful interactions occur as a result of insults expressed on social net-working websites. These comments foster an unprofessional tone on the internet, which is usually un-derstood and mitigated through passive mechanisms and techniques. Additionally, the recall rates of current systems that combine insult detection with machine learning and natural language processing are incredibly poor. To establish a viable classification scheme for such concepts, the research ana-lyzes how to identify bullying in writing by examining and testing various approaches. We propose an effective method to assess bullying, identify aggressive comments, and analyze their veracity. NLP and machine learning are employed to examine social perception and identify the aggressive impact on in-dividuals or groups. The ideal prototyping system for identifying cyber dangers in social media relies heavily on an efficient classifier. The goal of the paper is to emphasize the critical role that learning strategies play in enhancing natural language processing efficiency.