Quantifying and localizing state uncertainty in hidden Markov models using conditional entropy profiles

Durand Jean-Baptiste, Guédon Yann. 2014. Quantifying and localizing state uncertainty in hidden Markov models using conditional entropy profiles. In : Proceedings of COMPSTAT 2014, August 19-22, 2014, Geneva, Switzerland. Manfred Gilli, Gil Gonzalez-Rodriguez, Alicia Nieto-Reyes (eds.) - Genève : Université de Genève, 2014. s.l. : s.n., 213-221. ISBN 978-2-8399-1347-8 International Conference on Computational Statistics. 21, Genève, Suisse, 19 Août 2014/22 Août 2014.

Résumé : A family of graphical hidden Markov models that generalizes hidden Markov chain (HMC) and tree (HMT) models is introduced. It is shown that global uncertainty on the state process can be decomposed as a sum of conditional entropies that arte interpreted as local contributions to global uncertainty. An efficient algorithm is derived to compute conditional entropy profiles in the case of HMC and HMT models. The relevance of these profiles and their complementary with other state restoration algorithms for interpretation and diagnosis of hidden states is highlighted. It is also shown that classical smoothing profiles (posterior marginal probabilities of the states at each time, given the observations) cannot be related to global state uncertainty in the general case.

Classification Agris : U10 - Informatique, mathématiques et statistiques

Auteurs et affiliations

• Durand Jean-Baptiste, INRIA (FRA) ORCID: 0000-0001-6800-1438
• Guédon Yann, CIRAD-BIOS-UMR AGAP (FRA)

Source : Cirad - Agritrop (https://agritrop.cirad.fr/574646/)

Métriques

[ Page générée et mise en cache le 2023-06-22 ]