CoxSE: Exploring the Potential of Self-Explaining Neural Networks with Cox Proportional Hazards Model for Survival Analysis

TL;DR

This paper introduces CoxSE, a self-explaining neural network for survival analysis that maintains interpretability while achieving competitive predictive performance, and also proposes a hybrid model for improved robustness.

Contribution

The paper presents CoxSE, a locally explainable neural network for Cox models, and CoxSENAM, a hybrid model controlling explanation stability, advancing interpretability in survival analysis.

Findings

CoxSE provides stable, locally-linear explanations with high predictive accuracy.

CoxSENAM enhances explanation stability and model robustness.

NAM-based models are more robust to non-informative features.

Abstract

The Cox Proportional Hazards (CPH) model has long been the preferred survival model for its explainability. However, to increase its predictive power beyond its linear log-risk, it was extended to utilize deep neural networks, sacrificing its explainability. In this work, we explore the potential of self-explaining neural networks (SENN) for survival analysis. We propose a new locally explainable Cox proportional hazards model, named CoxSE, by estimating a locally-linear log-hazard function using the SENN. We also propose a modification to the Neural additive (NAM) model, hybrid with SENN, named CoxSENAM, which enables the control of the stability and consistency of the generated explanations. Several experiments using synthetic and real datasets are presented, benchmarking CoxSE and CoxSENAM against a NAM-based model, a DeepSurv model explained with SHAP, and a linear CPH model. The results show that, unlike the NAM-based model, the SENN-based model can provide more stable and consistent explanations while maintaining the predictive power of the black-box model. The results also show that, due to their structural design, NAM-based models demonstrate better robustness to non-informative features. Among the models, the hybrid model exhibits the best robustness.