A Study on Cyberbullying Detection Techniques in Online Social Networks

As the technological advancement, digital/social media consumption is increasing daily. Because social media sites like Twitter, Facebook, Instagram, Telegram, YouTube, and WhatsApp are so widely used, people may now engage more often, effectively, and efficiently. Cyberbullying is a sort of bullying that occurs on a digital platform with the purpose to harm and impair someone’s mental and emotional health. Cyberbullying is defined as bullying that occurs when a victim is utilizing a computer, a phone, a tablet, a social networking site, text messaging, a chat room, or other electronic device. Cyberbullying, which has become one of the most dangerous online threats for children, has caused considerable worries in the community due to the rapid growth in social media use.The proposed research work aims to design, firstly to study and analyze different issues and challenges faced by the people while using social media platforms. Secondly to develop a conventional classification framework for cyberbullying identification and prevention using machine learning. Developing a model to help in the prevention of text-based, image-based, and audio-based cyberbullying in the context of different social media platforms using a different language. The performance can be analyzed using different parameter like Accuracy, Precision, F1-score, ERR, Recall etc. Thirdly, it proposes identify and mitigate attacks based on online social media platforms. At last, to find the influence of the feature optimization or reduction on proposed classification framework.