SD - Discrimination
Framework for Subgroup Discovery, Discriminaton Detection and Exceptional Model Mining in datasets written in Python.
Subgroup discovery
- WRA (Weighed Relative Accuracy) measure
- Specificity measure
- Sensitivity measure
- Chi-squared measure
Discrimination detection
- Elift measure
- Support to Direct and Indirect discrimination
Exceptional Model Mining (EMM)
Model class
- Phi-Coefficient
Quality measure
- SCD (Significance of Correlation Difference)
Other features
- Implementation of Beam Search for automatic discovery