|
Activity Number:
|
37
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #305173 |
|
Title:
|
Efficient Bayesian Estimation for a General Dynamic Mixture Model
|
|
Author(s):
|
Christopher K. Carter*+ and Robert J. Kohn and Paolo Giordani
|
|
Companies:
|
University of New South Wales and University of New South Wales and Sveriges Riksbank
|
|
Address:
|
, , ,
|
|
Keywords:
|
State space models ; Mixture models ; nonparametric regression ; Markov chain Monte Carlo ; time series
|
|
Abstract:
|
A Bayesian approach is presented for estimating a mixture of linear Gaussian state space models where the mixture probabilities depend on the history of the series. Such models are suitable for interventions in time series and nonparametric regression. They are a generalization of the models considered in Gerlach, Carter and Kohn (JASA, 2000) where the mixture probabilities were assumed to follow a Markov chain. An efficient Markov chain Monte Carlo sampling scheme is derived for the general model and is shown to rapidly converge to the posterior distribution.
|
