Feature screening in ultrahigh dimensional Cox's model

Guangren Yang, Ye Yu, Runze Li, Anne Buu

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Survival data with ultrahigh dimensional covariates, such as genetic markers, have been collected in medical studies and other fields. In this work, we propose a feature screening procedure for the Cox model with ultrahigh dimensional covariates. The proposed procedure is distinguished from existing sure independence screening (SIS) procedures (Fan, Feng, and Wu (2010); Zhao and Li (2012)) in that it is based on the joint likelihood of potential active predictors, and therefore is not a marginal screening procedure. The proposed procedure can effectively identify active predictors that are jointly dependent but marginally independent of the response without performing an iterative procedure. We develop a computationally effective algorithm to carry it out and establish its ascent property. We further prove that the proposed procedure possesses the sure screening property: with probability tending to one, the selected variable set includes the actual active predictors. We conducted Monte Carlo simulation to evaluate the finite sample performance of the proposed procedure and compare it with existing SIS procedures. The proposed methodology is also demonstrated through an empirical analysis of a data example.

Original languageEnglish (US)
Pages (from-to)881-901
Number of pages21
JournalStatistica Sinica
Volume26
Issue number3
DOIs
StatePublished - Jul 2016

Fingerprint

Cox Model
Screening
Predictors
Covariates
Ascent
Cox model
Survival Data
Empirical Analysis
Iterative Procedure
Likelihood
Monte Carlo Simulation
Methodology
Dependent
Evaluate

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Yang, Guangren ; Yu, Ye ; Li, Runze ; Buu, Anne. / Feature screening in ultrahigh dimensional Cox's model. In: Statistica Sinica. 2016 ; Vol. 26, No. 3. pp. 881-901.
@article{931817bf483c421ca8c93515ca1013eb,
title = "Feature screening in ultrahigh dimensional Cox's model",
abstract = "Survival data with ultrahigh dimensional covariates, such as genetic markers, have been collected in medical studies and other fields. In this work, we propose a feature screening procedure for the Cox model with ultrahigh dimensional covariates. The proposed procedure is distinguished from existing sure independence screening (SIS) procedures (Fan, Feng, and Wu (2010); Zhao and Li (2012)) in that it is based on the joint likelihood of potential active predictors, and therefore is not a marginal screening procedure. The proposed procedure can effectively identify active predictors that are jointly dependent but marginally independent of the response without performing an iterative procedure. We develop a computationally effective algorithm to carry it out and establish its ascent property. We further prove that the proposed procedure possesses the sure screening property: with probability tending to one, the selected variable set includes the actual active predictors. We conducted Monte Carlo simulation to evaluate the finite sample performance of the proposed procedure and compare it with existing SIS procedures. The proposed methodology is also demonstrated through an empirical analysis of a data example.",
author = "Guangren Yang and Ye Yu and Runze Li and Anne Buu",
year = "2016",
month = "7",
doi = "10.5705/ss.2014.171",
language = "English (US)",
volume = "26",
pages = "881--901",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "3",

}

Feature screening in ultrahigh dimensional Cox's model. / Yang, Guangren; Yu, Ye; Li, Runze; Buu, Anne.

In: Statistica Sinica, Vol. 26, No. 3, 07.2016, p. 881-901.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Feature screening in ultrahigh dimensional Cox's model

AU - Yang, Guangren

AU - Yu, Ye

AU - Li, Runze

AU - Buu, Anne

PY - 2016/7

Y1 - 2016/7

N2 - Survival data with ultrahigh dimensional covariates, such as genetic markers, have been collected in medical studies and other fields. In this work, we propose a feature screening procedure for the Cox model with ultrahigh dimensional covariates. The proposed procedure is distinguished from existing sure independence screening (SIS) procedures (Fan, Feng, and Wu (2010); Zhao and Li (2012)) in that it is based on the joint likelihood of potential active predictors, and therefore is not a marginal screening procedure. The proposed procedure can effectively identify active predictors that are jointly dependent but marginally independent of the response without performing an iterative procedure. We develop a computationally effective algorithm to carry it out and establish its ascent property. We further prove that the proposed procedure possesses the sure screening property: with probability tending to one, the selected variable set includes the actual active predictors. We conducted Monte Carlo simulation to evaluate the finite sample performance of the proposed procedure and compare it with existing SIS procedures. The proposed methodology is also demonstrated through an empirical analysis of a data example.

AB - Survival data with ultrahigh dimensional covariates, such as genetic markers, have been collected in medical studies and other fields. In this work, we propose a feature screening procedure for the Cox model with ultrahigh dimensional covariates. The proposed procedure is distinguished from existing sure independence screening (SIS) procedures (Fan, Feng, and Wu (2010); Zhao and Li (2012)) in that it is based on the joint likelihood of potential active predictors, and therefore is not a marginal screening procedure. The proposed procedure can effectively identify active predictors that are jointly dependent but marginally independent of the response without performing an iterative procedure. We develop a computationally effective algorithm to carry it out and establish its ascent property. We further prove that the proposed procedure possesses the sure screening property: with probability tending to one, the selected variable set includes the actual active predictors. We conducted Monte Carlo simulation to evaluate the finite sample performance of the proposed procedure and compare it with existing SIS procedures. The proposed methodology is also demonstrated through an empirical analysis of a data example.

UR - http://www.scopus.com/inward/record.url?scp=85011802316&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011802316&partnerID=8YFLogxK

U2 - 10.5705/ss.2014.171

DO - 10.5705/ss.2014.171

M3 - Article

AN - SCOPUS:85011802316

VL - 26

SP - 881

EP - 901

JO - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 3

ER -