Multiple Attribute Fairness: Application to Fraud Detection

TL;DR

This paper introduces a new fairness measure and a model-agnostic heuristic for achieving multi-attribute fairness in classification, particularly applied to fraud detection, handling multiple protected attributes and high arity values.

Contribution

It proposes a novel fairness measure and an iterative heuristic that extends fairness to multiple attributes and high arity values, improving upon existing two-group fairness methods.

Findings

Achieves fairness across multiple protected attribute values.

Performs comparably to existing fairness techniques on public datasets.

Handles high arity attribute values effectively.

Abstract

We propose a fairness measure relaxing the equality conditions in the popular equal odds fairness regime for classification. We design an iterative, model-agnostic, grid-based heuristic that calibrates the outcomes per sensitive attribute value to conform to the measure. The heuristic is designed to handle high arity attribute values and performs a per attribute sanitization of outcomes across different protected attribute values. We also extend our heuristic for multiple attributes. Highlighting our motivating application, fraud detection, we show that the proposed heuristic is able to achieve fairness across multiple values of a single protected attribute, multiple protected attributes. When compared to current fairness techniques, that focus on two groups, we achieve comparable performance across several public data sets.