|
Activity Number:
|
194
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #302796 |
|
Title:
|
Likelihood-Based Sufficient Dimension Reduction for Regression
|
|
Author(s):
|
R. Dennis Cook*+
|
|
Companies:
|
The University of Minnesota
|
|
Address:
|
School of Statistics, Minneapolis, MN, 55455,
|
|
Keywords:
|
Central subspace ; Grassmann manifolds ; Principal fitted components
|
|
Abstract:
|
Dimension reduction is an old idea that has moved to a position of prominence in recent years because technological advances now allow scientists to routinely formulate regressions in which the number p of predictors is considerably larger than in the past. Starting with a definition of sufficient reductions, we will consider a variety of models for reducing the dimension of the predictor vector. The models start from one in which maximum likelihood estimation produces principal components, step along a few incremental expansions, and end with models that yield versatile methodology. Penalized log likelihoods will be suggested for isolating important predictors and for estimation in regressions with n < p. It will be argued that a likelihood-based approach to dimension reduction provides robust methodology that is superior to past methods.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
