IRJET - Cyberbulling Detection Model

This document describes a cyberbullying detection model that uses machine learning techniques to overcome limitations of existing methods. It analyzes a Twitter dataset containing annotated tweets using natural language processing and classifiers like SVM, random forest, and KNN. The models achieved up to 95% accuracy in detecting cyberbullying posts. The authors propose expanding the model to use unsupervised learning, integrate with social media APIs to detect bullying in real-time, and develop image recognition to identify bullying across multiple media platforms.