We are provided with a dataset consisting of social media comments and their manual classification into Toxic, Severe Toxic, Obscene, Threat, Insult, Identity hate and Non-Toxic. The task is to create multi-class toxic comment classifier using:
-
Different vectorization approaches studied in class
-
Using the avg-AUC (Area under the ROC) metric to measure the efficacy of different models
-
Hyperparameter tune the models
