|
Activity Number:
|
310
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #304899 |
|
Title:
|
A Nonparametric Plug-in Rule for Smoothing Parameter Selection
|
|
Author(s):
|
Soumendra N. Lahiri*+
|
|
Companies:
|
Iowa State University
|
|
Address:
|
315 E. Snedecor Hall, Ames, IA, 50011,
|
|
Keywords:
|
block bootstrap ; optimal block size ; smoothing parameter ; plug-in rule
|
|
Abstract:
|
In this talk, we describe a nonparametric plug-in principle for selecting smoothing parameters in nonparametric curve estimation problems using suitable resampling methods. The key idea is to combine the bootstrap and other resampling methods suitably so the various population parameters appearing in the theoretical optimal value of the smoothing parameter are estimated implicitly (i.e., without explicit analytical expressions). This proposed method possesses the computational simplicity of a plug-in approach but without the analytical work on the part of the user. Usefulness and properties of the proposed method are illustrated in problems involving optimal block length selection for block bootstrap and optimal bandwidth selection for nonparametric regression function/density estimation.
|
- 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 |
