Application of Spike and Slab Prior Elastic Network Cox Model in Cancer
10.11783/j.issn.1002-3674.2025.05.009
- VernacularTitle:Spike-and-slab先验弹性网络Cox模型在癌症中的应用
- Author:
Yue SU
1
;
Fudong WEN
1
;
Dan LIU
1
Author Information
1. 哈尔滨医科大学公共卫生学院卫生统计教研室(150081)
- Publication Type:Journal Article
- Keywords:
Bayesian statistics;
Spike and slab prior;
Elastic network;
Cox model;
Cancer
- From:
Chinese Journal of Health Statistics
2025;42(5):689-693
- CountryChina
- Language:Chinese
-
Abstract:
Objective To establish a high-precision and highly interpretable predictive model to address the challenges faced by constructing predictive models for high-dimensional omics data,such as many potential predictive factors,limited sample size,and high correlation between predictive factors.Methods Unify spike and slab priors and elastic network penalties into the Cox model and propose the spike and slab priors elastic network Cox model.This model can shrink the coefficients to varying degrees based on the importance of different variables.Use the expectation-maximization algorithm coordinate descent algorithm to fit the model,which estimates parameters by maximizing a posterior probability within a Bayesian framework.Results In different simulation experimental scenarios,the spike and slab prior elastic network Cox model exhibited higher sensitivity,balance accuracy,and concordance index than traditional models.In the validation analysis of real datasets,the concordance index of this model is also higher than that of traditional models.Conclusion The spike and slab prior elastic network Cox model is a new variable screening and survival prediction method that can handle high-dimensional omics data in cancer research.
