Estimation of Viterbi path in Bayesian hidden Markov models. / Lember, Jüri; Gasbarra, Dario; Koloydenko, Alexey; Kuljus, Kristi.

In: Metron, Vol. 77, No. 2 , 08.2019, p. 137-169.

Research output: Contribution to journalArticle

Published

Standard

Estimation of Viterbi path in Bayesian hidden Markov models. / Lember, Jüri; Gasbarra, Dario; Koloydenko, Alexey; Kuljus, Kristi.

In: Metron, Vol. 77, No. 2 , 08.2019, p. 137-169.

Research output: Contribution to journalArticle

Harvard

Lember, J, Gasbarra, D, Koloydenko, A & Kuljus, K 2019, 'Estimation of Viterbi path in Bayesian hidden Markov models', Metron, vol. 77, no. 2 , pp. 137-169. https://doi.org/10.1007/s40300-019-00152-7

APA

Vancouver

Author

Lember, Jüri ; Gasbarra, Dario ; Koloydenko, Alexey ; Kuljus, Kristi. / Estimation of Viterbi path in Bayesian hidden Markov models. In: Metron. 2019 ; Vol. 77, No. 2 . pp. 137-169.

BibTeX

@article{e478cbd7a0cd48b4a4c79e758a538ff0,
title = "Estimation of Viterbi path in Bayesian hidden Markov models",
abstract = "The article studies different methods for estimating the Viterbi path in the Bayesian framework. The Viterbi path is an estimate of the underlying state path in hidden Markov models (HMMs), which has a maximum posterior probability (MAP). For an HMM with given parameters, the Viterbi path can be easily found with the Viterbi algorithm. In the Bayesian framework the Viterbi algorithm is not applicable and several iterative methods can be used instead. We introduce a new EM-type algorithm for finding the MAP path and compare it with various other methods for finding the MAP path, including the variational Bayes approach and MCMC methods. Examples with simulated data are used to compare the performance of the methods. The main focus is on non-stochastic iterative methods and our results show that the best of those methods work as well or better than the best MCMC methods. Our results demonstrate that when the primary goal is segmentation, then it is more reasonable to perform segmentation directly by considering the transition and emission parameters as nuisance parameters.",
author = "J{\"u}ri Lember and Dario Gasbarra and Alexey Koloydenko and Kristi Kuljus",
year = "2019",
month = "8",
doi = "10.1007/s40300-019-00152-7",
language = "English",
volume = "77",
pages = "137--169",
journal = "Metron",
issn = "0026-1424",
publisher = "Springer",
number = "2",

}

RIS

TY - JOUR

T1 - Estimation of Viterbi path in Bayesian hidden Markov models

AU - Lember, Jüri

AU - Gasbarra, Dario

AU - Koloydenko, Alexey

AU - Kuljus, Kristi

PY - 2019/8

Y1 - 2019/8

N2 - The article studies different methods for estimating the Viterbi path in the Bayesian framework. The Viterbi path is an estimate of the underlying state path in hidden Markov models (HMMs), which has a maximum posterior probability (MAP). For an HMM with given parameters, the Viterbi path can be easily found with the Viterbi algorithm. In the Bayesian framework the Viterbi algorithm is not applicable and several iterative methods can be used instead. We introduce a new EM-type algorithm for finding the MAP path and compare it with various other methods for finding the MAP path, including the variational Bayes approach and MCMC methods. Examples with simulated data are used to compare the performance of the methods. The main focus is on non-stochastic iterative methods and our results show that the best of those methods work as well or better than the best MCMC methods. Our results demonstrate that when the primary goal is segmentation, then it is more reasonable to perform segmentation directly by considering the transition and emission parameters as nuisance parameters.

AB - The article studies different methods for estimating the Viterbi path in the Bayesian framework. The Viterbi path is an estimate of the underlying state path in hidden Markov models (HMMs), which has a maximum posterior probability (MAP). For an HMM with given parameters, the Viterbi path can be easily found with the Viterbi algorithm. In the Bayesian framework the Viterbi algorithm is not applicable and several iterative methods can be used instead. We introduce a new EM-type algorithm for finding the MAP path and compare it with various other methods for finding the MAP path, including the variational Bayes approach and MCMC methods. Examples with simulated data are used to compare the performance of the methods. The main focus is on non-stochastic iterative methods and our results show that the best of those methods work as well or better than the best MCMC methods. Our results demonstrate that when the primary goal is segmentation, then it is more reasonable to perform segmentation directly by considering the transition and emission parameters as nuisance parameters.

UR - https://arxiv.org/abs/1802.01630

U2 - 10.1007/s40300-019-00152-7

DO - 10.1007/s40300-019-00152-7

M3 - Article

VL - 77

SP - 137

EP - 169

JO - Metron

JF - Metron

SN - 0026-1424

IS - 2

ER -