|
Activity Number:
|
375
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #305089 |
|
Title:
|
Optimal Combining of Regression Procedures
|
|
Author(s):
|
Chihche Lin*+
|
|
Companies:
|
Astellas Pharma Global Development, Inc.
|
|
Address:
|
3 Parkway North, Deerfield, IL, 60015,
|
|
Keywords:
|
Adaptive Regression by Mixing ; Model Combining ; Model Averaging ; Model Selection ; ARM ; Optimal Combining
|
|
Abstract:
|
When a set of candidate models or estimation procedures is present with the intention to produce a highly adaptive accurate estimator, to reduce uncertainty due to model selection, model combining is an attractive alternative. The ARM (adaptive regression by mixing) method proposed by Yang (2001) to combine regression models leads to optimal rate of convergence offered by the candidate models. In this work, we pursue an asymptotic optimality of ARM by establishing a risk bound that shows ARM not only achieves the best rate of convergence over the candidates but can also attain the optimal constant asymptotically.
|
- 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 |
