Article |
A group bridge approach for variable selection
Department of Statistics and Actuarial Science, University of Iowa, 221 Schaeffer Hall, Iowa City, Iowa 52242, U.S.A. jian-huang{at}uiowa.edu
Division of Biostatistics, Department of Epidemiology and Public Health, Yale University, New Haven, Connecticut 06520, U.S.A. shuangge.ma{at}yale.edu
Department of Management Science, University of Miami, Coral Gables, Florida 33124, U.S.A. hxie{at}exchange.sba.miami.edu
Department of Statistics, Rutgers University, Piscataway, New Jersey 08854, U.S.A. cunhui{at}stat.rutgers.edu
Received for publication 1 May 2007.
Revision received 1 October 2008.
 |
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 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.
Key Words: Bridge estimator • Iterative lasso • Penalized regression • Two-level selection • Variable-selection consistency