Cyberbullying-and-Hatespeech-Detection

This was a year long project where my team and I worked on online hatespeech detection. The Kaggle Dataset "Dataset for detection of cyber trolls" was used. Preprocessing includes tokenization, Porter's Stemming Algorithm and Cosine distance. We have created a neural network to detect cyberbullying. It's performance was compared to multiple classifiers and a pre trained BERT model. We also attempted to integrate our model with Telegram for real time Hatespeech detection. You can read more about this project in this IEEE Paper