Published Paper Details:

A significant societal issue, particularly among youngsters, is cyberbullying. "The deliberate, repeated, and hostile use of information technology to harm or harass other people" is the definition of cyberbullying. It has increased in popularity since the creation of social media platforms like Twitter and Facebook. The Internet is the sole foundation of contemporary society. People today have a hard time imagining life without the internet. People have been utilizing social networking sites in recent years to exchange information, thoughts, and ideas. Different types of content, including text, images, audio, and video data, may be used in these exchanges. Giving users the opportunity to manage the messages posted on their own private space is a significant problem in today's online social networks (OSNs), since it prevents the broadcast of inappropriate information. As of right present, OSNs hardly support this need. As a result, social media academics are placing more and more emphasis on the automatic identification of messages that engage in cyberbullying. The use of rules and norms, human moderators, and blacklists based on offensive language are some of the traditional strategies for combating cyberbullying. In order to automatically identify cyberbullying behaviors, a principled learning framework must be created. A flexible rule-based system that enables users to tailor the filtering criteria to be factual to their walls and a soft classifier powered by machine learning that automatically labels messages in content-based filtering help achieve this. Server can learn the words and save them in the database based on this filtering. Before the message can be exchanged, the server might review the words at the moment of transmission and prohibit.