Model selection for Cox models with time-varying coefficients

Jun Yan; Jian Huang

doi:10.1111/j.1541-0420.2011.01692.x

Back

Model selection for Cox models with time-varying coefficients

Journal article

Peer reviewed

Model selection for Cox models with time-varying coefficients

Jun Yan and Jian Huang

Biometrics, Vol.68(2), pp.419-428

06/2012

DOI: 10.1111/j.1541-0420.2011.01692.x

PMCID: PMC3384767

PMID: 22506825

View Online

Abstract

Summary Cox models with time-varying coefficients offer great flexibility in capturing the temporal dynamics of covariate effects on right-censored failure times. Because not all covariate coefficients are time varying, model selection for such models presents an additional challenge, which is to distinguish covariates with time-varying coefficient from those with time-independent coefficient. We propose an adaptive group lasso method that not only selects important variables but also selects between time-independent and time-varying specifications of their presence in the model. Each covariate effect is partitioned into a time-independent part and a time-varying part, the latter of which is characterized by a group of coefficients of basis splines without intercept. Model selection and estimation are carried out through a fast, iterative group shooting algorithm. Our approach is shown to have good properties in a simulation study that mimics realistic situations with up to 20 variables. A real example illustrates the utility of the method.

Algorithms

Data Interpretation, Statistical

Time Factors

Computer Simulation

Humans

Proportional Hazards Models

Survival Analysis

Linear Models

Biometry - methods

Liver Cirrhosis, Biliary - mortality

Details

Title: Subtitle: Model selection for Cox models with time-varying coefficients
Creators: Jun Yan - Department of Statistics, University of Connecticut, Storrs, Connecticut 06269, USA. jun.yan@uconn.edu
Jian Huang
Resource Type: Journal article
Publication Details: Biometrics, Vol.68(2), pp.419-428
DOI: 10.1111/j.1541-0420.2011.01692.x
PMID: 22506825
PMCID: PMC3384767
NLM abbreviation: Biometrics
ISSN: 1541-0420
eISSN: 1541-0420
Publisher: United States
Grant note: R01 CA142774 / NCI NIH HHS R01 CA120988-04 / NCI NIH HHS R01 CA120988 / NCI NIH HHS R01CA142774 / NCI NIH HHS R01CA120988 / NCI NIH HHS
Language: English
Date published: 06/2012
Academic Unit: Statistics and Actuarial Science
Record Identifier: 9983985714902771

Metrics

63 Record Views

41 Times Cited - Web of Science