Grouped Variables Independence Screening in Sparse Ultra High-Dimensional Feature Space

• Activity Number: 49
• Type: Invited
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Rui Song*+
• Companies: Colorado State University
• Keywords: independent learning ; sure independence screening ; variable selection

Abstract

In this work, we extend the correlation learning to marginal grouped variable screening. Our grouped variable independence screening is called GIS, a specific member of the sure independence screening. Several closely related variable screening procedures are proposed. It is shown that under some mild technical conditions, the proposed independence screening methods enjoy a sure screening property. The extent to which the dimensionality can be reduced by independence screening is also explicitly quantified. This is a joint work with Jianqing Fan and Yang Feng.