maddlib

This repository contains the source code of the Python package related to the MADD metric.

Installation

This package is available on pypi and GitHub.

It can be installed using pip (with command line pip install maddlib or python3 -m pip install maddlib).

If an error occurs, please check first that your Python version is greater than 3.10.4.

Getting started

Some example notebooks are in tutorials folder. You can use the MADD to evaluate algorithmic fairness or to mitigate algorithmic unfairness.

You can find the full documentation here: https://melinaverger.github.io/documentation-maddlib/.

Reference

If you are using this package, please cite:

M. Verger, S. Lallé, F. Bouchet, and V. Luengo. Is Your Model ”MADD”? A Novel Metric to Evaluate Algorithmic Fairness for Predictive Student Models. In M. Feng, T. Käser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Data Mining, pages 91–102, Bengaluru, India, July 2023. International Educational Data Mining Society.

@inproceedings{Verger2023,  
    title = {Is Your Model ``MADD''? A Novel Metric to Evaluate Algorithmic Fairness for Predictive Student Models},  
    author = {M\'{e}lina Verger and S\'{e}bastien Lall\'{e} and Fran\c{c}ois Bouchet and Vanda Luengo},  
    booktitle = {Proceedings of the 16th International Conference on Educational Data Mining},  
    editor = {Mingyu Feng and Tanja Käser and Partha Talukdar},  
    doi = {10.5281/zenodo.8115786},  
    isbn = {978-1-7336736-4-8},  
    month = {July},  
    address = {Bengaluru, India},  
    pages = {91--102},  
    publisher = {International Educational Data Mining Society},  
    year = {2023}  
}

License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.