|
Activity Number:
|
564
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #303621 |
|
Title:
|
Sparse Sufficient Dimension Reduction and Variable Selection
|
|
Author(s):
|
Xin Chen*+ and R. Dennis Cook
|
|
Companies:
|
University of Minnesota and The University of Minnesota
|
|
Address:
|
313 Ford Hall 224 Church Street S.E., Minneapolis, MN, 55455,
|
|
Keywords:
|
sufficient dimension reduction ; variable selection ; Grassmann manifolds ; principal components ; lasso
|
|
Abstract:
|
Sufficient dimension reduction (SDR) means the construction of a few derived variables from original variables without loss of information. SDR is very helpful especially when the number of original variables is large. However each derived variable usually consists of a linear combination of all original variables, making it difficult to interpret. It is desirable to reduce the dimensionality sufficiently and select important variables at the same time. To achieve this goal, we propose a subspace oriented method that incorporates a coordinate-independent penalty term to a series of model-based and model-free SDR approaches.
|
