|
Activity Number:
|
410
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #304177 |
|
Title:
|
Fully Bayesian Smoothing Splines for Varying-Coefficient Models
|
|
Author(s):
|
Yu Yue*+
|
|
Companies:
|
Baruch College
|
|
Address:
|
Department of Statistics and CIS, New York, NY, 10010,
|
|
Keywords:
|
|
|
Abstract:
|
Abstract: Smoothing spline is one of the most widely used penalized estimates in statistics. Wahba (1978) introduced an exact Bayesian version of smoothing spline using limiting Gaussian prior. Speckman and Sun (2003) proved Wahba's limiting prior belongs to a class of priors termed "partially informative normal (PIN)" priors. In this work, we show how to apply PIN priors to the varying-coefficient models. We provide sensible noninformative priors on variance parameters for fully Bayesian inference. In addition, a Bayes factor approach is developed for testing nonlinear effects in the model. The method is demonstrated through a simulation study and an application in a time-dependent data set.
|
- 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 |
