Cyberbullying Detection Using Machine Learning

Abstract

Cyberbullying has recently been reported to cause significant damage to society and the economy. Technological advances regarding the annotation of web documents and the diversity of online communities make the detection and tracking of such cases quite difficult and very difficult. Cyberbullying occurs through various information and communication technologies such as instant messaging, email, text messaging, social networking sites, blogs, and websites. We must fight this kind of abuse. The purpose of this project is to analyze and evaluate the communication between two specific individuals or an anonymous person. Recommended machine learning model to detect and prevent harassment in one-on-one conversation. The classifiers, SVM, Logistic Regression, Random Forest and Gradient Boosting are used to train and test social media bullying.  We use the Twitter chat dataset to train the ml model.