Density-Tempered Marginalized Sequential Monte Carlo Samplers

33 Pages Posted: 12 May 2011 Last revised: 18 Oct 2013

See all articles by Jin-Chuan Duan

Jin-Chuan Duan

National University of Singapore (NUS) - Business School and Risk Management Institute

Andras Fulop

ESSEC Business School

Date Written: October 17, 2013

Abstract

We propose a density-tempered marginalized sequential Monte Carlo (SMC) sampler, a new class of samplers for full Bayesian inference of general state-space models. The dynamic states are approximately marginalized out using a particle filter, and the parameters are sampled via a sequential Monte Carlo sampler over a density-tempered bridge between the prior and the posterior. Our approach delivers exact draws from the joint posterior of the parameters and the latent states for any given number of state particles, and thus easily parallelizable in implementation. We also build into the proposed method a device that can automatically select a suitable number of state particles. Since the method incorporates sample information in a smooth fashion, it delivers good performance in the presence of outliers. We check the performance of the density-tempered SMC algorithm using simulated data based on a linear Gaussian state space model with and without mis-specification. We also apply it on real stock prices using a GARCH-type model with microstructure noise.

Keywords: Particle Filter, MCMC, Sequential Monte Carlo Samplers, Bayesian Methods

JEL Classification: C11

Suggested Citation

Duan, Jin-Chuan and Fulop, Andras, Density-Tempered Marginalized Sequential Monte Carlo Samplers (October 17, 2013). Available at SSRN: https://ssrn.com/abstract=1837772 or http://dx.doi.org/10.2139/ssrn.1837772

Jin-Chuan Duan

National University of Singapore (NUS) - Business School and Risk Management Institute ( email )

1 Business Link
Singapore, 117592
Singapore

Andras Fulop (Contact Author)

ESSEC Business School ( email )

3 Avenue Bernard Hirsch
CS 50105 CERGY
CERGY, CERGY PONTOISE CEDEX 95021
France

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
306
Abstract Views
1,534
Rank
180,971
PlumX Metrics