fairsd 0.1.0

Fair Subgroup Discovery fork of the original FairSD by Maurizio Pulizzi

FairSD

FairSD is a package that implements top-k subgroup discovery algorithms for identifying subgroups that may be treated unfairly by a machine learning model.

The package has been designed to offer the user the possibility to use different notions of fairness as quality measures. Integration with the Fairlearn package allows the user to use all the fairlearn metrics as quality measures. The user can also define custom quality measures, by extending the QualityFunction class present in qualitymeasures.py module.

Acknowledgements

This package is a fork of the original fairsd repository by Maurizio Pulizzi. The original repository can be found here. However, since the original repository is no longer maintained, my fork is used to add convenience features and bug fixes and thus will be released in this package.

Acknowledgements to the original repository

Some parts of the code are an adaptation of the pysubgroup package. These parts are indicated in the code.

Contributors to the original repository

• Maurizio Pulizzi
• Hilde Weerts