Fare auditing diagnostics and pairwise fairness error metrics for ranking.
This repository contains code and example analysis for evaluating the fairness of rankings with respect to protected groups, using the pairwise error metrics and auditing methodology presented in the paper:
“FARE: Diagnostics for Fair Ranking using Pairwise Error Metrics” in the proceedings of the Web Conference (WWW 2019) by Caitlin Kuhlman, MaryAnn VanValkenburg, Elke Rundensteiner
This work is released under the 3-Clause BSD License.
The three pairwise error metrics presented in the paper, Rank Equality, Rank Parity, and Rank Calibration are included in the fare package distibution, along with methods to perform fairness auditing of rankings.
Example analysis, including the experiments in the paper, is available in the jupyter notebooks in the examples folder.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.