Survival Analysis with Graph-Based Regularization for Predictors

TL;DR

This paper introduces a graph-regularized Cox model for survival analysis that effectively incorporates variable correlations, improving variable selection accuracy and robustness in biomedical datasets.

Contribution

It proposes a novel graph-based regularization method for Cox models, with an efficient algorithm and theoretical guarantees, enhancing variable selection in survival analysis.

Findings

Outperforms existing methods in synthetic datasets

Demonstrates improved accuracy in organ transplantation data

Provides theoretical recovery guarantees

Abstract

We study the variable selection problem in survival analysis to identify the most important factors affecting survival time. Our method incorporates prior knowledge of mutual correlations among variables, represented through a graph. We utilize the Cox proportional hazard model with a graph-based regularizer for variable selection. We present a computationally efficient algorithm developed to solve the graph regularized maximum likelihood problem by establishing connections with the group lasso, and provide theoretical guarantees about the recovery error and asymptotic distribution of the proposed estimators. The improved performance of the proposed approach compared with existing methods are demonstrated in both synthetic and real organ transplantation datasets.