Towards Model-Agnostic Post-Hoc Adjustment for Balancing Ranking Fairness and Algorithm Utility

Introduction

The implementation of xorder This repo includes source code and data files of experiments on benchmark datasets.

Requirements

• Python 3.x
• pytorch 1.5.0
• scikit-learn 0.22.1
• numpy 1.18.1
• pandas 1.0.1
• numba 0.48.0

Run the experiments

The experiment result with logistic regression classifier and xauc disparity metric on compas can be obtained with:
python3 run_experiment.py --dataset compas --classifier lr --eval_metric xauc

Selections of datasets, ranking fairness metrics and classifiers.

--dataset: adult, compas, framingham, german
--eval_metric: xauc, prf
--classifier: lr(logistic regression), rb(bipatite rankboost)