Counterfactual explanation of machine learning survival models

Kovalev, M.; Utkin, L.; Coolen, F.; Konstantinov, A.

Abstract

A method for counterfactual explanation of machine learning survival models is proposed. One of the difficulties of solving the counterfactual explanation problem is that the classes of examples are implicitly defined through outcomes of a machine learning survival model in the form of survival functions. A condition that establishes the difference between survival functions of the original example and the counterfactual is introduced. This condition is based on using a distance between mean times to event. It is shown that the counterfactual explanation problem can be reduced to a standard convex optimization problem with linear constraints when the explained black-box model is the Cox model. For other black-box models, it is proposed to apply the wellknown Particle Swarm Optimization algorithm. Numerical experiments with real and synthetic data demonstrate the proposed method.

Citation

Kovalev, M., Utkin, L., Coolen, F., & Konstantinov, A. (2022). Counterfactual explanation of machine learning survival models. Informatica: An International Journal, 32(4), 817-847. https://doi.org/10.15388/21-infor468

• Journal Article Type: Article
• Acceptance Date: Dec 2, 2021
• Online Publication Date: Dec 9, 2021
• Publication Date: 2022
• Deposit Date: Dec 4, 2021
• Publicly Available Date: Dec 6, 2021
• Journal: Informatica
• Print ISSN: 0868-4952
• Electronic ISSN: 1822-8844
• Publisher: Vilnius University
• Peer Reviewed: Peer Reviewed
• Volume: 32
• Issue: 4
• Pages: 817-847
• DOI: https://doi.org/10.15388/21-infor468
• Keywords: interpretable model; explainable AI; survival analysis; censored data; convex optimization; counterfactual explanation; Cox model; Particle Swarm Optimization
• Public URL: https://durham-repository.worktribe.com/output/1221271