Adaptive-Cox model averaging for right-censored data

Yu Mei Chang, Pao Sheng Shen, Chun Shu Chen

Research output: Contribution to journalArticle

Abstract

In medical studies, Cox proportional hazards model is a commonly used method to deal with the right-censored survival data accompanied by many explanatory covariates. In practice, the Akaike's information criterion (AIC) or the Bayesian information criterion (BIC) is usually used to select an appropriate subset of covariates. It is well known that neither the AIC criterion nor the BIC criterion dominates for all situations. In this paper, we propose an adaptive-Cox model averaging procedure to get a more robust hazard estimator. First, by applying AIC and BIC criteria to perturbed datasets, we obtain two model averaging (MA) estimated survival curves, called AIC-MA and BIC-MA. Then, based on Kullback–Leibler loss, a better estimate of survival curve between AIC-MA and BIC-MA is chosen, which results in an adaptive-Cox estimate of survival curve. Simulation results show the superiority of our approach and an application of the proposed method is also presented by analyzing the German Breast Cancer Study dataset.

Original languageEnglish
Pages (from-to)9364-9376
Number of pages13
JournalCommunications in Statistics - Theory and Methods
Volume46
Issue number19
DOIs
Publication statusPublished - 2017 Oct 2

Fingerprint

Model Averaging
Cox Model
Right-censored Data
Bayesian Information Criterion
Akaike Information Criterion
Curve
Covariates
Censored Survival Data
Cox Proportional Hazards Model
Breast Cancer
Hazard
Estimate
Estimator
Subset
Simulation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

Chang, Yu Mei ; Shen, Pao Sheng ; Chen, Chun Shu. / Adaptive-Cox model averaging for right-censored data. In: Communications in Statistics - Theory and Methods. 2017 ; Vol. 46, No. 19. pp. 9364-9376.
@article{dc80271da8694ccdbbdc0da4c8441f3c,
title = "Adaptive-Cox model averaging for right-censored data",
abstract = "In medical studies, Cox proportional hazards model is a commonly used method to deal with the right-censored survival data accompanied by many explanatory covariates. In practice, the Akaike's information criterion (AIC) or the Bayesian information criterion (BIC) is usually used to select an appropriate subset of covariates. It is well known that neither the AIC criterion nor the BIC criterion dominates for all situations. In this paper, we propose an adaptive-Cox model averaging procedure to get a more robust hazard estimator. First, by applying AIC and BIC criteria to perturbed datasets, we obtain two model averaging (MA) estimated survival curves, called AIC-MA and BIC-MA. Then, based on Kullback–Leibler loss, a better estimate of survival curve between AIC-MA and BIC-MA is chosen, which results in an adaptive-Cox estimate of survival curve. Simulation results show the superiority of our approach and an application of the proposed method is also presented by analyzing the German Breast Cancer Study dataset.",
author = "Chang, {Yu Mei} and Shen, {Pao Sheng} and Chen, {Chun Shu}",
year = "2017",
month = "10",
day = "2",
doi = "10.1080/03610926.2016.1208237",
language = "English",
volume = "46",
pages = "9364--9376",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "19",

}

Adaptive-Cox model averaging for right-censored data. / Chang, Yu Mei; Shen, Pao Sheng; Chen, Chun Shu.

In: Communications in Statistics - Theory and Methods, Vol. 46, No. 19, 02.10.2017, p. 9364-9376.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Adaptive-Cox model averaging for right-censored data

AU - Chang, Yu Mei

AU - Shen, Pao Sheng

AU - Chen, Chun Shu

PY - 2017/10/2

Y1 - 2017/10/2

N2 - In medical studies, Cox proportional hazards model is a commonly used method to deal with the right-censored survival data accompanied by many explanatory covariates. In practice, the Akaike's information criterion (AIC) or the Bayesian information criterion (BIC) is usually used to select an appropriate subset of covariates. It is well known that neither the AIC criterion nor the BIC criterion dominates for all situations. In this paper, we propose an adaptive-Cox model averaging procedure to get a more robust hazard estimator. First, by applying AIC and BIC criteria to perturbed datasets, we obtain two model averaging (MA) estimated survival curves, called AIC-MA and BIC-MA. Then, based on Kullback–Leibler loss, a better estimate of survival curve between AIC-MA and BIC-MA is chosen, which results in an adaptive-Cox estimate of survival curve. Simulation results show the superiority of our approach and an application of the proposed method is also presented by analyzing the German Breast Cancer Study dataset.

AB - In medical studies, Cox proportional hazards model is a commonly used method to deal with the right-censored survival data accompanied by many explanatory covariates. In practice, the Akaike's information criterion (AIC) or the Bayesian information criterion (BIC) is usually used to select an appropriate subset of covariates. It is well known that neither the AIC criterion nor the BIC criterion dominates for all situations. In this paper, we propose an adaptive-Cox model averaging procedure to get a more robust hazard estimator. First, by applying AIC and BIC criteria to perturbed datasets, we obtain two model averaging (MA) estimated survival curves, called AIC-MA and BIC-MA. Then, based on Kullback–Leibler loss, a better estimate of survival curve between AIC-MA and BIC-MA is chosen, which results in an adaptive-Cox estimate of survival curve. Simulation results show the superiority of our approach and an application of the proposed method is also presented by analyzing the German Breast Cancer Study dataset.

UR - http://www.scopus.com/inward/record.url?scp=85020225453&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020225453&partnerID=8YFLogxK

U2 - 10.1080/03610926.2016.1208237

DO - 10.1080/03610926.2016.1208237

M3 - Article

AN - SCOPUS:85020225453

VL - 46

SP - 9364

EP - 9376

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 19

ER -