Hierarchical Bayesian Survival Analysis and Projective Covariate Selection in Cardiovascular Event Risk Prediction
Abstract
Identifying biomarkers with predictive value for disease risk stratification is an important task in epidemiology. This paper describes an application of Bayesian linear survival regres-sion to model cardiovascular event risk in di-abetic individuals with measurements avail-able on 55 candidate biomarkers. We extend the survival model to include data from a larger set of non-diabetic individuals in an e↵ort to increase the predictive performance for the diabetic subpopulation. We com-pare the Gaussian, Laplace and horseshoe shrinkage priors, and find that the last has the best predictive performance and shrinks strong predictors less than the others. We implement the projection predictive covari-ate selection approach of Dupuis and Robert (2003) to further search for small sets of pre-dictive biomarkers that could provide cost-ecient prediction without significant loss in performance. In passing, we present a deriva-tion of the projective covariate selection in Bayesian decision theoretic framework. 1
Cite
Text
Peltola et al. "Hierarchical Bayesian Survival Analysis and Projective Covariate Selection in Cardiovascular Event Risk Prediction." Conference on Uncertainty in Artificial Intelligence, 2014.Markdown
[Peltola et al. "Hierarchical Bayesian Survival Analysis and Projective Covariate Selection in Cardiovascular Event Risk Prediction." Conference on Uncertainty in Artificial Intelligence, 2014.](https://mlanthology.org/uai/2014/peltola2014uai-hierarchical/)BibTeX
@inproceedings{peltola2014uai-hierarchical,
title = {{Hierarchical Bayesian Survival Analysis and Projective Covariate Selection in Cardiovascular Event Risk Prediction}},
author = {Peltola, Tomi and Havulinna, Aki S. and Salomaa, Veikko and Vehtari, Aki},
booktitle = {Conference on Uncertainty in Artificial Intelligence},
year = {2014},
pages = {79-88},
url = {https://mlanthology.org/uai/2014/peltola2014uai-hierarchical/}
}