|
Activity Number:
|
129
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #305331 |
|
Title:
|
A New Class of Model Selection Strategies with Applications in Mixed Model Selection
|
|
Author(s):
|
Jiming Jiang*+
|
|
Companies:
|
University of California, Davis
|
|
Address:
|
1 Shields Ave., Davis, CA, 95616-5270,
|
|
Keywords:
|
|
|
Abstract:
|
Many model search strategies involve trading off model fit with model complexity in a penalized goodness-of-fit measure. Such strategies do not always translate into good finite sample performance, as the procedure is typically overly sensitive to the setting of penalty parameters. Furthermore, these strategies do not generalize naturally to more complex models and nonconventional problems, in which the penalties and model complexity may not be defined naturally. In this talk, we introduce a new class of model selection strategies that are potentially less subjective and applicable to both conventional and nonconventional problems. We apply these strategies to linear and generalized linear mixed model selection problems.
|
- 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 2006 program |
