HarX: Real-time harassment detection tool using machine learning

Abstract

Cybersecurity has a great deal of importance over the digital market for organizations in this modern era. Nowadays all kinds of communications and connections are established by using the internet. Chatting is a main source of communication. The major problem faced by users is harassment. User starts to get harassed frequently as most of users does not know what to do and how to take action or how to stop this. In this work, we employ machine learning and natural language processing to tackle online harassment. This study proposed a real time machine learning based algorithm which detects harassment actively and alert user to take action against it. For detection mechanism, Naïve Bayes classification is used. The proposed approach attain approximately 77% accuracy. The result shows that the algorithm actively detects harassing keywords in chat messages.