Exact bayesian prediction in non-gaussian Markov-switching model - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2009

Exact bayesian prediction in non-gaussian Markov-switching model

Résumé

In this paper we consider a class of recently introduced jump-Markov switching models, involving a hidden process X, an observed process Y and a latent process R which models the switches or changes of regimes in (X,Y). We address the Bayesian prediction problem, and we show that the p-step ahead a posteriori conditional expectation (and associated conditional covariance matrix) can be computed exactly linearly in time.
Fichier non déposé

Dates et versions

hal-01367823 , version 1 (16-09-2016)

Identifiants

  • HAL Id : hal-01367823 , version 1

Citer

Noémie Bardel, François Desbouvries. Exact bayesian prediction in non-gaussian Markov-switching model. ASMDA 2009 : XIIIth International Conference on Applied Stochastic Models and Data Analysis, Jun 2009, Vilnius, Lithuania. pp.199 - 202. ⟨hal-01367823⟩
59 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More