Open Access
2007 An improved bound for the exponential stability of predictive filters of hidden Markov models
László Gerencser, György Michaletzky, Gábor Molnár-Sáska
Commun. Inf. Syst. 7(2): 133-152 (2007).

Abstract

We consider hidden Markov processes in discrete time with a finite state space $X$ and a general observation or read-out space $Y$, which is assumed to be a Polish space. It is well-known that in the statistical analysis of HMMs the so-called predictive filter plays a fundamental role. A useful result establishing the exponential stability of the predictive filter with respect to perturbations of its initial condition was given in "Exponential forgetting and geometric ergodicity in hidden Markov models" (F. LeGland, L. Mevel, Mathematics of control, signals and systems, 13(2000), pp. 63-93) in the case, when the assumed transition probability matrix was primitive. The main technical result of the present paper is the extension of the cited result by showing that the random constant and the deterministic positive exponent showing up in the inequality stating exponential stability can be chosen so that for any prescribed ${s\geq 0}$ the $s$-th exponential moment of the random constant is finite. An application of this result to the estimation of HMMs with primitive transition probabilities will be also briefly presented.

Citation

Download Citation

László Gerencser. György Michaletzky. Gábor Molnár-Sáska. "An improved bound for the exponential stability of predictive filters of hidden Markov models." Commun. Inf. Syst. 7 (2) 133 - 152, 2007.

Information

Published: 2007
First available in Project Euclid: 20 July 2007

zbMATH: 1263.62117
MathSciNet: MR2344193

Keywords: Doeblin-condition , Hidden Markov models , L-mixing , predictive filters , Random mappings , risk processes

Rights: Copyright © 2007 International Press of Boston

Vol.7 • No. 2 • 2007
Back to Top