Journal article
A Group Bridge Approach for Variable Selection
Biometrika, Vol.96(2), pp.339-355
06/01/2009
DOI: 10.1093/biomet/asp020
PMCID: PMC2796848
PMID: 20037673
Abstract
In multiple regression problems when covariates can be naturally grouped, it is important to carry out feature selection at the group and within-group individual-variable levels simultaneously. The existing methods, including the Lasso and group Lasso, are designed for either variable selection or group selection, but not for both. We propose a group bridge approach that it is capable of simultaneous selection at both the group and within-group individual variable levels. The proposed approach is a penalized regularization method that uses a specially designed group bridge penalty. It has the oracle group selection property, in that it can correctly select important groups with probability converging to one. In contrast, the group Lasso and group least angle regression methods in general do not possess such an oracle property in group selection. Simulation studies indicate that the group bridge has superior performance in group and individual variable selection relative to several existing methods.
Details
- Title: Subtitle
- A Group Bridge Approach for Variable Selection
- Creators
- JIAN HUANG - Department of Statistics and Actuarial Science, University of Iowa, 221 Schaeffer Hall, Iowa City, Iowa 52242, U.S.ASHUANGGE MA - Division of Biostatistics, Department of Epidemiology and Public Health, Yale University, New Haven, Connecticut 06520, U.S.AHUILIANG XIE - Department of Management Science, University of Miami, Coral Gables, Florida 33124, U.S.ACUN-HUI ZHANG - Department of Statistics, Rutgers University, Piscataway, New Jersey 08854, U.S.A
- Resource Type
- Journal article
- Publication Details
- Biometrika, Vol.96(2), pp.339-355
- DOI
- 10.1093/biomet/asp020
- PMID
- 20037673
- PMCID
- PMC2796848
- NLM abbreviation
- Biometrika
- ISSN
- 0006-3444
- eISSN
- 1464-3510
- Grant note
- DOI: 10.13039/100000002, name: U.S. National Institutes of Health; DOI: 10.13039/100000001, name: National Science Foundation; DOI: 10.13039/100009226, name: National Security Agency; DOI: 10.13039/100000002, name: NIH
- Language
- English
- Date published
- 06/01/2009
- Academic Unit
- Statistics and Actuarial Science
- Record Identifier
- 9983985980402771
Metrics
38 Record Views