|
Activity Number:
|
137
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #309740 |
|
Title:
|
Objective Priors for Spatially Adaptive Smoothing Splines
|
|
Author(s):
|
Yu Yue*+
|
|
Companies:
|
University of Missouri-Columbia
|
|
Address:
|
52 Broadway Village Dr, Columbia, MO, 65201,
|
|
Keywords:
|
|
|
Abstract:
|
In previous work, Yue and Speckman developed a class of priors for spatially adaptive smoothing and thin-plate splines by introducing a spatial model for the precision parameter of an intrinsic Gaussian Markov random field. The adaptive priors are effective in the classic nonparametric regression problem with normal error terms (i.e., curve fitting). The computation is efficient due to the sparseness of the precision matrix. In this work, we discuss the possibility of using objective priors on the variance terms of the adaptive priors. The results are mixed. Unlike the nonadaptive case investigated by Speckman and Sun (2003), it appears that improper priors cannot be used simultaneously at all levels of the hierarchical model. However, we find conditions on proper priors in the high levels of the model to permit the use of the invariance prior for the error variance.
|
- 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 |
