Cyber-Bullying Detection Via Text Mining and Machine Learning

Cyber Bullying is one of the most recent evils of social media. With a boom in the usage of social media, the freedom of expression is being exploited. Statistics show that overall 36.5 percent people think they have been cyberbullied in their lifetime. These numbers are more than double of what they were in 2007, and there is an increase from 2018–19, suggesting we are heading in the wrong direction. Solutions to curtail this issue to a certain extent have already been deployed in the market. However, they possess limitations of usage, or simply do not use efficient algorithms. This paper aims at identifying cyberbullying at is origin, meaning when it is being drafted in real-time. Identifying traces of cyberbullying before the content is uploaded on the internet can help reduce circulation of hurtful messages. Using Machine Learning with the support of Natural Language Processing(NLP, results in better performance of cyberbullying detection.