Parameter Learning and Change Detection Using a Particle Filter With Accelerated Adaptation

37 Pages Posted: 22 Jun 2018 Last revised: 13 Sep 2018

See all articles by Karol Gellert

Karol Gellert

University of Technology Sydney (UTS) - Quantitative Finance Research Centre

Erik Schlögl

University of Technology Sydney (UTS), Quantitative Finance Research Centre; University of Cape Town (UCT) - The African Institute of Financial Markets and Risk Management; Faculty of Science, Department of Statistics, University of Johannesburg; Financial Research Network (FIRN)

Date Written: June 8, 2018

Abstract

This paper presents the construction of a particle filter, which incorporates elements inspired by genetic algorithms, in order to achieve accelerated adaptation of the estimated posterior distribution to changes in model parameters. Specifically, the filter is designed for the situation where the subsequent data in online sequential filtering does not match the model posterior filtered based on data up to a current point in time. The examples considered encompass parameter regime shifts and stochastic volatility. The filter adapts to regime shifts extremely rapidly and delivers a clear heuristic for distinguishing between regime shifts and stochastic volatility, even though the model dynamics assumed by the filter exhibit neither of those features.

Keywords: Particle Filter, Estimation, Sequential Monte Carlo, Sequential Bayesian Updating, Regime Shift, Stochastic Volatility

JEL Classification: G10, C11, C13, C32, C53, C58

Suggested Citation

Gellert, Karol and Schloegl, Erik, Parameter Learning and Change Detection Using a Particle Filter With Accelerated Adaptation (June 8, 2018). FIRN Research Paper, Forthcoming, Available at SSRN: https://ssrn.com/abstract=3192709 or http://dx.doi.org/10.2139/ssrn.3192709

Karol Gellert

University of Technology Sydney (UTS) - Quantitative Finance Research Centre ( email )

P.O. Box 123
Sydney
Australia

Erik Schloegl (Contact Author)

University of Technology Sydney (UTS), Quantitative Finance Research Centre ( email )

Ultimo
PO Box 123
Sydney, NSW 2007
Australia
+61 2 9514 2535 (Phone)

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

University of Cape Town (UCT) - The African Institute of Financial Markets and Risk Management ( email )

Leslie Commerce Building
Rondebosch
Cape Town, Western Cape 7700
South Africa

Faculty of Science, Department of Statistics, University of Johannesburg ( email )

Auckland Park, 2006
South Africa

Financial Research Network (FIRN)

C/- University of Queensland Business School
St Lucia, 4071 Brisbane
Queensland
Australia

HOME PAGE: http://www.firn.org.au

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
59
Abstract Views
642
rank
482,994
PlumX Metrics