Publication

Abstract : With the dawn of the new age of communication, i.e., social media, the world has come into the palms of every individual at every corner of the world. However, easier access to messaging implies easier forms of abuse. As the usage of social media increases, the other side of communication, bullying, also increases. Cyberbullying refers to the concept of harassing individuals or groups through means of the internet with the help of messages or multimedia. Cyberbullying has a huge impact on mental health and research is ongoing in this area to accurately detect bullying text. Existing models are embedded into social media platforms to either ban or apply sanctions on the users involved in the passage of such materials. The proposed study aims to develop a model with a similar approach to detect cyberbullying in comment text. Also, sentiment analysis is performed on the comment text to get an idea of the message’s context, thus improving the chances of bullying detection. The work is divided into two stages. In the first stage, the text and the sentiment score are directly passed to the machine learning models. Later, during the second stage word embedding using Word2Vec, a natural language processing technique, is performed on the comment text to obtain vectorized features which are then given to the machine learning models along with sentiment features. It was observed that the latter model shows high levels of success and that it can be deployed in real-world applications.