|
Activity Number:
|
452
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #307816 |
|
Title:
|
Adaptive Independent Metropolis-Hastings Using Normal Mixtures
|
|
Author(s):
|
Robert Kohn*+ and Paolo Giordani
|
|
Companies:
|
University of New South Wales and Bank of Sweden
|
|
Address:
|
Sydney 2052 , NSW, Sydney , International, 2052, Australia
|
|
Keywords:
|
Clustering ; Markov chain Monte Carlo ; Semiparametric regression models ; State space models.
|
|
Abstract:
|
Adaptive Metropolis-Hastings samplers use information obtained from previous draws to tune the proposal distribution. The tuning is carried out automatically, often repeatedly, and continues after the burn-in period. We propose a method for adaptive independent Metropolis-Hastings sampling using a mixture of normals as a proposal distribution. To take full advantage of the potential of adaptive sampling our algorithm updates the mixture of normals frequently, starting early in the chain. The algorithm is built for speed and reliability and its sampling performance is evaluated with simulated examples and with applications to time-varying-parameter, semi-parametric, and stochastic volatility models.
|
