|
Activity Number:
|
226
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
IMS
|
| Abstract - #307802 |
|
Title:
|
Localized Model Selection by Cross-Validation
|
|
Author(s):
|
Yuhong Yang*+
|
|
Companies:
|
The University of Minnesota
|
|
Address:
|
313 Ford Hall, School of Statistics, Minneapolis, MN, 55455,
|
|
Keywords:
|
model selection ; localized model selection ; cross validation
|
|
Abstract:
|
Traditional model selection theory and practice focus on selecting a single one that is deemed the best in a global sense. In high dimensional or complex data situations, when very distinct learning procedures are considered, it is often the case that the candidates perform differently in ranking in different regions. This motivates localized model selection that allows more flexibility to improve over the global model selection strategy. In this talk, we will focus on localized cross validation. Our goal is to select adaptively the procedure that performs the best locally among the candidate regression procedures. Theoretical and empirical results will be provided.
|
- 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 2007 program |
