|
Activity Number:
|
416
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #306685 |
|
Title:
|
Bayesian Circular Regression
|
|
Author(s):
|
Barbara Jane George*+ and Kaushik Ghosh
|
|
Companies:
|
U.S. Environmental Protection Agency and New Jersey Institute of Technology
|
|
Address:
|
109 TW Alexander Drive, Research Triangle Park, NC, 27709,
|
|
Keywords:
|
directional data ; MCMC ; predictive density ; Dirichlet process
|
|
Abstract:
|
Circular data, where observations are represented by points on the unit circle, arise in the modeling of diverse phenomenon such as time of peak blood glucose level or direction of ocean waves. In this talk, we discuss a semiparametric Bayesian approach for regression involving circular variables. The proposed approach uses Markov chain Monte Carlo techniques to fit a flexible regression model. We will illustrate the proposed method with an example.
|
