Machine Learning Approaches for Cyberbullying Detection in Social Media

Bullying has been an enduring issue throughout human history, with its manifestations evolving over time.The strategies have evolved in tandem with technological progress, shifting from conventional physical bullying to the contemporary issue of cyber bullying. The digital age, marked by the proliferation of user-generated web content, especially on social media platforms, has witnessed a disturbing surge in hate speech. This research project undertakes a systematic review of published studies focusing on approaches to detect cyberbullying and explores methods for identifying hate speech on social media. Additionally, the project conducts a comprehensive comparative study of various supervised machine learning algorithms, encompassing both standard and ensemble methods, to assess their effectiveness in tackling this pressing issue. The ultimate aim is to contribute to the mitigation of hate speech online and its ominous real-world consequences, including violence towards minorities, mass shootings, lynching’s, and ethnic cleansing.