Variable selection via the weighted group lasso for factor analysis models

Kei Hirose, Sadanori Konishi

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

We consider the problem of selecting variables in factor analysis models. The L 1 regularization procedure is introduced to perform an automatic variable selection. In the factor analysis model, each variable is controlled by multiple factors when there are more than one underlying factor. We treat parameters corresponding to the multiple factors as grouped parameters, and then apply the group lasso. Furthermore, the weight of the group lasso penalty is modified to obtain appropriate estimates and improve the performance of variable selection. Crucial issues in this modeling procedure include the selection of the number of factors and a regularization parameter. Choosing these parameters can be viewed as a model selection and evaluation problem. We derive a model selection criterion for evaluating the factor analysis model via the weighted group lasso. Monte Carlo simulations are conducted to investigate the effectiveness of the proposed procedure. A real data example is also given to illustrate our procedure.

Original languageEnglish
Pages (from-to)345-361
Number of pages17
JournalCanadian Journal of Statistics
Volume40
Issue number2
DOIs
Publication statusPublished - Jun 1 2012
Externally publishedYes

Fingerprint

Lasso
Variable Selection
Factor Analysis
Selection of Variables
Model Evaluation
Model Selection Criteria
Regularization Parameter
Model Selection
Model
Penalty
Regularization
Monte Carlo Simulation
Factor analysis
Variable selection
Factors
Modeling
Estimate

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Variable selection via the weighted group lasso for factor analysis models. / Hirose, Kei; Konishi, Sadanori.

In: Canadian Journal of Statistics, Vol. 40, No. 2, 01.06.2012, p. 345-361.

Research output: Contribution to journalArticle

@article{850af7a05faf4b75b01aac67f4fd21ec,
title = "Variable selection via the weighted group lasso for factor analysis models",
abstract = "We consider the problem of selecting variables in factor analysis models. The L 1 regularization procedure is introduced to perform an automatic variable selection. In the factor analysis model, each variable is controlled by multiple factors when there are more than one underlying factor. We treat parameters corresponding to the multiple factors as grouped parameters, and then apply the group lasso. Furthermore, the weight of the group lasso penalty is modified to obtain appropriate estimates and improve the performance of variable selection. Crucial issues in this modeling procedure include the selection of the number of factors and a regularization parameter. Choosing these parameters can be viewed as a model selection and evaluation problem. We derive a model selection criterion for evaluating the factor analysis model via the weighted group lasso. Monte Carlo simulations are conducted to investigate the effectiveness of the proposed procedure. A real data example is also given to illustrate our procedure.",
author = "Kei Hirose and Sadanori Konishi",
year = "2012",
month = "6",
day = "1",
doi = "10.1002/cjs.11129",
language = "English",
volume = "40",
pages = "345--361",
journal = "Canadian Journal of Statistics",
issn = "0319-5724",
publisher = "Statistical Society of Canada",
number = "2",

}

TY - JOUR

T1 - Variable selection via the weighted group lasso for factor analysis models

AU - Hirose, Kei

AU - Konishi, Sadanori

PY - 2012/6/1

Y1 - 2012/6/1

N2 - We consider the problem of selecting variables in factor analysis models. The L 1 regularization procedure is introduced to perform an automatic variable selection. In the factor analysis model, each variable is controlled by multiple factors when there are more than one underlying factor. We treat parameters corresponding to the multiple factors as grouped parameters, and then apply the group lasso. Furthermore, the weight of the group lasso penalty is modified to obtain appropriate estimates and improve the performance of variable selection. Crucial issues in this modeling procedure include the selection of the number of factors and a regularization parameter. Choosing these parameters can be viewed as a model selection and evaluation problem. We derive a model selection criterion for evaluating the factor analysis model via the weighted group lasso. Monte Carlo simulations are conducted to investigate the effectiveness of the proposed procedure. A real data example is also given to illustrate our procedure.

AB - We consider the problem of selecting variables in factor analysis models. The L 1 regularization procedure is introduced to perform an automatic variable selection. In the factor analysis model, each variable is controlled by multiple factors when there are more than one underlying factor. We treat parameters corresponding to the multiple factors as grouped parameters, and then apply the group lasso. Furthermore, the weight of the group lasso penalty is modified to obtain appropriate estimates and improve the performance of variable selection. Crucial issues in this modeling procedure include the selection of the number of factors and a regularization parameter. Choosing these parameters can be viewed as a model selection and evaluation problem. We derive a model selection criterion for evaluating the factor analysis model via the weighted group lasso. Monte Carlo simulations are conducted to investigate the effectiveness of the proposed procedure. A real data example is also given to illustrate our procedure.

UR - http://www.scopus.com/inward/record.url?scp=84861211116&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84861211116&partnerID=8YFLogxK

U2 - 10.1002/cjs.11129

DO - 10.1002/cjs.11129

M3 - Article

AN - SCOPUS:84861211116

VL - 40

SP - 345

EP - 361

JO - Canadian Journal of Statistics

JF - Canadian Journal of Statistics

SN - 0319-5724

IS - 2

ER -