Hate Speech Detection on Twitter

This project is in collaboration with Zehao Su and Stefan Thoma finished in Jan 2021.

To summarize:

  1. I conducted researches on classifying tweets into hate speech, offensive speech, or neither.
  2. I implemented data imputation method like SMOTE to overcome the data imbalance and performs several experiments including BERT-Transformer and SVM to disentangle the semantic space.