Explainable Machine Learning for Fraud Detection

TL;DR

This paper investigates explainability techniques for machine learning models used in real-time fraud detection, focusing on background dataset selection and runtime trade-offs for both supervised and unsupervised methods.

Contribution

It provides insights into optimizing explainability approaches specifically tailored for fraud detection systems, addressing practical implementation challenges.

Findings

Background dataset selection impacts explanation quality.

Runtime trade-offs are crucial for real-time applications.

Supervised and unsupervised models require different explainability strategies.

Abstract

The application of machine learning to support the processing of large datasets holds promise in many industries, including financial services. However, practical issues for the full adoption of machine learning remain with the focus being on understanding and being able to explain the decisions and predictions made by complex models. In this paper, we explore explainability methods in the domain of real-time fraud detection by investigating the selection of appropriate background datasets and runtime trade-offs on both supervised and unsupervised models.