[Filtrage exact dans les modèles conditionnellement linéaires à sauts markoviens]
Dans les modèles classiques linéaires cachés à sauts markoviens, le calcul exact des filtres optimaux est dʼune complexité exponentielle par rapport au nombre dʼobservations, ce qui nécessite lʼutilisation de techniques dʼapproximation. Dans cet article, nous introduisons une nouvelle famille de modèles qui rendent ces opérations possibles en pratique.
In the classical setup of Markov switching hidden linear systems, the exact evaluation of optimal filters requires computations with complexity increasing exponentially with respect to the number of observations. In the present article, we propose a new family of models which overcome this difficulty, and render practically feasible these calculations.
Accepté le :
Publié le :
@article{CRMATH_2011__349_9-10_587_0,
author = {Pieczynski, Wojciech},
title = {Exact filtering in conditionally {Markov} switching hidden linear models},
journal = {Comptes Rendus. Math\'ematique},
pages = {587--590},
publisher = {Elsevier},
volume = {349},
number = {9-10},
year = {2011},
doi = {10.1016/j.crma.2011.02.007},
language = {en},
url = {http://www.numdam.org/articles/10.1016/j.crma.2011.02.007/}
}
TY - JOUR AU - Pieczynski, Wojciech TI - Exact filtering in conditionally Markov switching hidden linear models JO - Comptes Rendus. Mathématique PY - 2011 SP - 587 EP - 590 VL - 349 IS - 9-10 PB - Elsevier UR - http://www.numdam.org/articles/10.1016/j.crma.2011.02.007/ DO - 10.1016/j.crma.2011.02.007 LA - en ID - CRMATH_2011__349_9-10_587_0 ER -
%0 Journal Article %A Pieczynski, Wojciech %T Exact filtering in conditionally Markov switching hidden linear models %J Comptes Rendus. Mathématique %D 2011 %P 587-590 %V 349 %N 9-10 %I Elsevier %U http://www.numdam.org/articles/10.1016/j.crma.2011.02.007/ %R 10.1016/j.crma.2011.02.007 %G en %F CRMATH_2011__349_9-10_587_0
Pieczynski, Wojciech. Exact filtering in conditionally Markov switching hidden linear models. Comptes Rendus. Mathématique, Tome 349 (2011) no. 9-10, pp. 587-590. doi : 10.1016/j.crma.2011.02.007. http://www.numdam.org/articles/10.1016/j.crma.2011.02.007/
[1] N. Abbassi, W. Pieczynski, Exact filtering in semi-Markov jumping system, in: Sixth International Conference of Computational Methods in Sciences and Engineering, Hersonissos, Crete, Greece, September 25–30, 2008.
[2] Efficient particle filtering for jump Markov systems. Application to time-varying autoregressions, IEEE Trans. Signal Process., Volume 51 (2003) no. 7, pp. 1762-1770
[3] Inference in Hidden Markov Models, Springer, 2005
[4] Discrete Time Markov Jump Linear Systems, Springer-Verlag, New York, 2005
[5] Signal and image segmentation using pairwise Markov chains, IEEE Trans. Signal Process., Volume 52 (2004) no. 9, pp. 2477-2489
[6] A unified approach to nonlinearity, structural change, and outliers, J. Econometrics, Volume 137 (2007), pp. 112-133
[7] Multisensor triplet Markov chains and theory of evidence, Internat. J. Approx. Reason., Volume 45 (2007) no. 1, pp. 1-16
[8] W. Pieczynski, Exact calculation of optimal filter in semi-Markov switching model, in: Fourth World Conference of the International Association for Statistical Computing (IASC 2008), Yokohama, Japan, December 5–8, 2008.
[9] Adaptive estimation and identification for discrete systems with Markov jump parameters, IEEE Trans. Automat. Control, Volume AC-25 (1982), pp. 1054-1065
[10] Deterministic approximate inference techniques for conditionally Gaussian state space models, Statist. Comput., Volume 16 (2006), pp. 279-292
Cité par Sources :
