harshilpatel1799 / Resampling-Techs-for-Imbalanced-Data

Resampling-Techs-for-Imbalanced-Data

Python imbalanced-learn Package

Illustration for Resampling Methods

Undersampling Techniques:

1. Random Undersampling
2. Tomek Links
3. ENN (Edited Nearest Neighbours)

Oversampling Techniques:

1. Random Oversampling
2. SMOTE (Synthetic Minority Over-Sampling Technique)
3. ADASYN (Adaptive Synthetic Sampling)

Combined Resampling:

1. SMOTETomek
2. SMOTEENN

Case Study of Imbalance Data: Bank Marketing

Tidy Data:

1. Check missing
2. Get X and y
3. Check imbalance
4. Transformation on X and y

The Metric Trap

Confusion Matrix

Apply Resampling

Recommand Readings