Cyber-Bullying Detection in Social Media Platform using Machine Learning

Now a day’s our smart gadgets are not only devices but true friends of human-being. Social-Networking, one from them provides us a virtual home far from home, where everyone feels connected even from thousand miles is one of the brighter sides of new era. The dark side of this coin is equally the worst, as this also increases the vulnerability of young people to threatening situations online.This Paper is divided into three main tasks, as a very first task, we explored various forms of Cyber-Crime, reviewed Cyber-Bullying, its forms, methods, effects, and the available recent research to detect and prevent it. Secondly, for the experimental purpose, we have collected data of Twitter’s 35000+ tweets, prepared/wrangled that data to fed it to various smart machine learning algorithms, then applied five important ML algorithms to those tweets for classification and prediction into two main classes ‘offensive’ or ‘non-offensive’. Finally, a comparison has been done among those ML algorithms based on several performance metrics.