detecting_bullies

Using scikit-learn to find bullies

Link to blogpost where I explain in more detail my approach.

In this folder, I tackle Kaggle's (very old) Detecting Insults in Social Commentary challenge.

Using Word Embeddings, I better the AUC ROC score from 0.84 to 0.89.