An exploration of techniques to deal with unbalanced classes in classification problems.
Four notebooks are provided here:
- General considerations - set up toy data, run some classifers to get a baseline
- Class weight adjustment
- Resampling: SMOTE and majority class downsampling
- Blagging