A Suite of Fairness Datasets for Tabular Classification

TL;DR

This paper introduces a comprehensive suite of 20 fairness datasets for tabular classification to facilitate more rigorous and standardized evaluation of fairness algorithms in machine learning.

Contribution

It provides a new collection of datasets and metadata to improve the consistency and depth of fairness research in tabular data classification.

Findings

Enables more thorough fairness algorithm testing

Standardizes dataset access for fairness research

Supports better benchmarking of fairness methods

Abstract

There have been many papers with algorithms for improving fairness of machine-learning classifiers for tabular data. Unfortunately, most use only very few datasets for their experimental evaluation. We introduce a suite of functions for fetching 20 fairness datasets and providing associated fairness metadata. Hopefully, these will lead to more rigorous experimental evaluations in future fairness-aware machine learning research.