imbalance-learning

There are some algorithm implementations for learning from imbalanced data, including over-sampling, under-sampling, and boosting.

1. Oversampling

• RandomOverSampling
• SMOTE
• Borderline-SMOTE1 and Borderline-SMOTE2
• ADASYN
• Safe-Level-SMOTE
• MWMOTE
• CGMOS

2. Undersampling

• RandomUnderSampling

3. Boosting

• SMOTEBoost
• EasyEnsemble
• BalanceCascade
• RUSBoost

P.S. Some of the implementations are rewritten or further optimized based on the original author's codes.