Variable selection for proportional odds model

Wenbin Lu, Hao Zhang

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

In this paper we study the problem of variable selection for the proportional odds model, which is a useful alternative to the proportional hazards model and might be appropriate when the proportional hazards assumption is not satisfied. We propose to fit the proportional odds model by maximizing the marginal likelihood subject to a shrinkage-type penalty, which encourages sparse solutions and hence facilitates the process of variable selection. Two types of shrinkage penalties are considered: the LASSO and the adaptive-LASSO (ALASSO) penalty. In the ALASSO penalty, different weights are imposed on different coefficients such that important variables are more protectively retained in the final model while unimportant ones are more likely to be shrunk to zeros. We further provide an efficient computation algorithm to implement the proposed methods, and demonstrate their performance through simulation studies and an application to real data. Numerical results indicate that both methods can produce accurate and interpretable models, and the ALASSO tends to work better than the usual LASSO.

Original languageEnglish (US)
Pages (from-to)3771-3781
Number of pages11
JournalStatistics in Medicine
Volume26
Issue number20
DOIs
StatePublished - Sep 10 2007
Externally publishedYes

Fingerprint

Proportional Odds Model
Adaptive Lasso
Variable Selection
Penalty
Selection of Variables
Shrinkage
Proportional Hazards Models
Weights and Measures
Marginal Likelihood
Proportional Hazards
Proportional Hazards Model
Likely
Simulation Study
Tend
Numerical Results
Alternatives
Zero
Coefficient
Model
Demonstrate

Keywords

  • Marginal likelihood
  • Proportional odds model
  • Shrinkage estimate
  • Variable selection

ASJC Scopus subject areas

  • Epidemiology

Cite this

Variable selection for proportional odds model. / Lu, Wenbin; Zhang, Hao.

In: Statistics in Medicine, Vol. 26, No. 20, 10.09.2007, p. 3771-3781.

Research output: Contribution to journalArticle

Lu, Wenbin ; Zhang, Hao. / Variable selection for proportional odds model. In: Statistics in Medicine. 2007 ; Vol. 26, No. 20. pp. 3771-3781.
@article{cc7a8b30dfd2441391cf9c175e972cf1,
title = "Variable selection for proportional odds model",
abstract = "In this paper we study the problem of variable selection for the proportional odds model, which is a useful alternative to the proportional hazards model and might be appropriate when the proportional hazards assumption is not satisfied. We propose to fit the proportional odds model by maximizing the marginal likelihood subject to a shrinkage-type penalty, which encourages sparse solutions and hence facilitates the process of variable selection. Two types of shrinkage penalties are considered: the LASSO and the adaptive-LASSO (ALASSO) penalty. In the ALASSO penalty, different weights are imposed on different coefficients such that important variables are more protectively retained in the final model while unimportant ones are more likely to be shrunk to zeros. We further provide an efficient computation algorithm to implement the proposed methods, and demonstrate their performance through simulation studies and an application to real data. Numerical results indicate that both methods can produce accurate and interpretable models, and the ALASSO tends to work better than the usual LASSO.",
keywords = "Marginal likelihood, Proportional odds model, Shrinkage estimate, Variable selection",
author = "Wenbin Lu and Hao Zhang",
year = "2007",
month = "9",
day = "10",
doi = "10.1002/sim.2833",
language = "English (US)",
volume = "26",
pages = "3771--3781",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "20",

}

TY - JOUR

T1 - Variable selection for proportional odds model

AU - Lu, Wenbin

AU - Zhang, Hao

PY - 2007/9/10

Y1 - 2007/9/10

N2 - In this paper we study the problem of variable selection for the proportional odds model, which is a useful alternative to the proportional hazards model and might be appropriate when the proportional hazards assumption is not satisfied. We propose to fit the proportional odds model by maximizing the marginal likelihood subject to a shrinkage-type penalty, which encourages sparse solutions and hence facilitates the process of variable selection. Two types of shrinkage penalties are considered: the LASSO and the adaptive-LASSO (ALASSO) penalty. In the ALASSO penalty, different weights are imposed on different coefficients such that important variables are more protectively retained in the final model while unimportant ones are more likely to be shrunk to zeros. We further provide an efficient computation algorithm to implement the proposed methods, and demonstrate their performance through simulation studies and an application to real data. Numerical results indicate that both methods can produce accurate and interpretable models, and the ALASSO tends to work better than the usual LASSO.

AB - In this paper we study the problem of variable selection for the proportional odds model, which is a useful alternative to the proportional hazards model and might be appropriate when the proportional hazards assumption is not satisfied. We propose to fit the proportional odds model by maximizing the marginal likelihood subject to a shrinkage-type penalty, which encourages sparse solutions and hence facilitates the process of variable selection. Two types of shrinkage penalties are considered: the LASSO and the adaptive-LASSO (ALASSO) penalty. In the ALASSO penalty, different weights are imposed on different coefficients such that important variables are more protectively retained in the final model while unimportant ones are more likely to be shrunk to zeros. We further provide an efficient computation algorithm to implement the proposed methods, and demonstrate their performance through simulation studies and an application to real data. Numerical results indicate that both methods can produce accurate and interpretable models, and the ALASSO tends to work better than the usual LASSO.

KW - Marginal likelihood

KW - Proportional odds model

KW - Shrinkage estimate

KW - Variable selection

UR - http://www.scopus.com/inward/record.url?scp=34548170943&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34548170943&partnerID=8YFLogxK

U2 - 10.1002/sim.2833

DO - 10.1002/sim.2833

M3 - Article

C2 - 17266170

AN - SCOPUS:34548170943

VL - 26

SP - 3771

EP - 3781

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 20

ER -