Efficient Learning via Simulation: A Marginalized Resample-Move Approach

50 Pages Posted: 12 Dec 2010 Last revised: 7 Mar 2013

See all articles by Andras Fulop

Andras Fulop

ESSEC Business School

Junye Li

ESSEC Business School

Date Written: December 29, 2012

Abstract

In state-space models, parameter learning is practically difficult and is still an open issue. This paper proposes an efficient simulation-based parameter learning method. First, the approach breaks up the interdependence of the hidden states and the static parameters by marginalizing out the states using a particle filter. Second, it applies a Bayesian resample-move approach to this marginalized system. The methodology is generic and needs little design effort. Different from batch estimation methods, it provides posterior quantities necessary for full sequential inference and recursive model monitoring. The algorithm is implemented both on simulated data in a linear Gaussian model for illustration and comparison and on real data in a Lévy jump stochastic volatility model and a structural credit risk model.

Keywords: State-Space Models, Particle Filters, Parameter Learning, State Filtering, Resample-Move, Markov Chain Monte Carlo, Stochastic Volatility, Credit Risk

JEL Classification: C11, C13, C32, G13

Suggested Citation

Fulop, Andras and Li, Junye, Efficient Learning via Simulation: A Marginalized Resample-Move Approach (December 29, 2012). Available at SSRN: https://ssrn.com/abstract=1724203 or http://dx.doi.org/10.2139/ssrn.1724203

Andras Fulop

ESSEC Business School ( email )

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

HOME PAGE: http://www.andrasfulop.com

Junye Li (Contact Author)

ESSEC Business School ( email )

5 Nepal Park
Singapore, Singapore 139408
Singapore

Register to save articles to
your library

Register

Paper statistics

Downloads
526
Abstract Views
2,369
rank
53,283
PlumX Metrics