Feature Selection Using General Index Model for Motif Screening

• Activity Number: 559
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: IMS
• Author(s): Wenxuan Zhong*+
• Companies: University of Illinois at Urbana-Champaign
• Keywords: dimension reduction ; feature selection ; motif discovery

Abstract

In this talk, a stepwise procedure will be introduced for variable selection under the sufficient dimension reduction framework, in which the response variable Y is influenced by the predictors through an unknown function of a few linear combinations of them. Unlike linear stepwise regression, the proposed method does not impose a special form of relationship (such as linear) between the response variable and the predictor variables. The proposed method will select variables that attain the maximum correlation between the transformed response and the linear combination of the variables by stepwisely add and delete a variable. The variable selection performance under diverging number of predictors and sample size has been investigated and will be discussed in the talk. The excellent empirical performance of the method will be demonstrated in the motif discovery analysis for mouse's embryonic development.