An Ensemble Learning Model for Automatic Detection of Cyberbullying on Instagram Platform

Social media usage has increased dramatically in recent years with the expansion of the Internet and also has emerged as the most powerful networking tool of the twenty-first century. However, increased social networking often has a detrimental influence on society, contributing to a number of unpleasant behaviors including online abuse, harassment, cyber-bullying, cyber-crime, as well as online trolling. Cyberbullying commonly causes severe emotional and physical anguish, especially among women as well as children, and may even lead to suicide attempts. Because of its significant detrimental societal effect, online harassment garners attention. Many occurrences have lately happened throughout the globe as a result of online bullying, such as the publishing of private conversations, rumors, and sexual comments. As a result, experts are increasingly interested in detecting bullying on social media via text or communication. The main purpose of this research is to create and refine a strategy for identifying cyber harassment and bullying comments by Machine Learning (ML) algorithms. In this paper, multi classifier system is termed an Ensemble-based learning model. The suggested model contains recommended ML algorithms including Support Vector Machine (SVM), Naive Bayes (NB), as well as Random Forest (RF) method as the classifier for this persistence. Finally Boosting Ensemble classifier has combined the best performance in detecting the abusing messages. The proposed model will be evaluated by various performance metrics.