Imbalanced Classification
Go to main | Course Page
Section 1: Introduction
- Problem with Imbalanced Classes
- Resources
- Build and examine baseline
- Download Resources
Section 2: Sampling Based Approaches
- Approach 1: Upsampling the minority class
- Approach 2: Downsampling the majority class
- Approach 3: Cluster Centroids - Theory
- Code Demo
- Approach 4: SMOTE Algorithm Theory
- Code Demo
- Approach 5: ADASYN Algorithm Theory
- Code Demo
- Approach 6: Near Miss
- Approach 7: Tomek Links
- Code Demo
Section 3: Class Sensitive Learning Approaches
- Class Sensitive Learning Theory
- Code Demo
- Class Weighted XGBoost and Demo
Section 4: Advanced Methods
- One Class Classification Theory
- One Class using SVM
- Mahalanobis Distance for One Class Classification
- Isolation Forest Demo
- Data Leakage - A sneaky surprisingly hard to detect example
- Edited Nearest Neighbour based example
- Probability Threshold Tuning