Recursive maximum likelihood estimation for hidden semi-Markov models

Kevin Squire, Stephen E Levinson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The term hidden semi-Markov model (HSMM) refers to a large class of stochastic models developed to address some of the shortcomings of hidden Markov models (HMMs). As with HMMs, the underlying sequence of states of a process is modelled as a discrete Markov chain. Unlike HMMs, each state in an HSMM can emit a variable length sequence of observations, with many ways to model duration and observation densities. Parameter estimation in HSMMs is typically done using EM or Viterbi (dynamic programming) algorithms. These algorithms require batch processing of large amounts of data, and so are not useful for online learning. To address this issue, we present here a recursive maximum-likelihood estimation (RMLE) algorithm for online estimation of HSMM parameters, based on a similar method developed for HMMs.

Original languageEnglish (US)
Title of host publication2005 IEEE Workshop on Machine Learning for Signal Processing
Pages329-334
Number of pages6
DOIs
StatePublished - Dec 1 2005
Event2005 IEEE Workshop on Machine Learning for Signal Processing - Mystic, CT, United States
Duration: Sep 28 2005Sep 30 2005

Publication series

Name2005 IEEE Workshop on Machine Learning for Signal Processing

Other

Other2005 IEEE Workshop on Machine Learning for Signal Processing
CountryUnited States
CityMystic, CT
Period9/28/059/30/05

Fingerprint

Maximum likelihood estimation
Hidden Markov models
Stochastic models
Dynamic programming
Parameter estimation
Markov processes

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Squire, K., & Levinson, S. E. (2005). Recursive maximum likelihood estimation for hidden semi-Markov models. In 2005 IEEE Workshop on Machine Learning for Signal Processing (pp. 329-334). [1532923] (2005 IEEE Workshop on Machine Learning for Signal Processing). https://doi.org/10.1109/MLSP.2005.1532923

Recursive maximum likelihood estimation for hidden semi-Markov models. / Squire, Kevin; Levinson, Stephen E.

2005 IEEE Workshop on Machine Learning for Signal Processing. 2005. p. 329-334 1532923 (2005 IEEE Workshop on Machine Learning for Signal Processing).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Squire, K & Levinson, SE 2005, Recursive maximum likelihood estimation for hidden semi-Markov models. in 2005 IEEE Workshop on Machine Learning for Signal Processing., 1532923, 2005 IEEE Workshop on Machine Learning for Signal Processing, pp. 329-334, 2005 IEEE Workshop on Machine Learning for Signal Processing, Mystic, CT, United States, 9/28/05. https://doi.org/10.1109/MLSP.2005.1532923
Squire K, Levinson SE. Recursive maximum likelihood estimation for hidden semi-Markov models. In 2005 IEEE Workshop on Machine Learning for Signal Processing. 2005. p. 329-334. 1532923. (2005 IEEE Workshop on Machine Learning for Signal Processing). https://doi.org/10.1109/MLSP.2005.1532923
Squire, Kevin ; Levinson, Stephen E. / Recursive maximum likelihood estimation for hidden semi-Markov models. 2005 IEEE Workshop on Machine Learning for Signal Processing. 2005. pp. 329-334 (2005 IEEE Workshop on Machine Learning for Signal Processing).
@inproceedings{975384d8e71c461ab7f0dca1aa5582fe,
title = "Recursive maximum likelihood estimation for hidden semi-Markov models",
abstract = "The term hidden semi-Markov model (HSMM) refers to a large class of stochastic models developed to address some of the shortcomings of hidden Markov models (HMMs). As with HMMs, the underlying sequence of states of a process is modelled as a discrete Markov chain. Unlike HMMs, each state in an HSMM can emit a variable length sequence of observations, with many ways to model duration and observation densities. Parameter estimation in HSMMs is typically done using EM or Viterbi (dynamic programming) algorithms. These algorithms require batch processing of large amounts of data, and so are not useful for online learning. To address this issue, we present here a recursive maximum-likelihood estimation (RMLE) algorithm for online estimation of HSMM parameters, based on a similar method developed for HMMs.",
author = "Kevin Squire and Levinson, {Stephen E}",
year = "2005",
month = "12",
day = "1",
doi = "10.1109/MLSP.2005.1532923",
language = "English (US)",
isbn = "0780395174",
series = "2005 IEEE Workshop on Machine Learning for Signal Processing",
pages = "329--334",
booktitle = "2005 IEEE Workshop on Machine Learning for Signal Processing",

}

TY - GEN

T1 - Recursive maximum likelihood estimation for hidden semi-Markov models

AU - Squire, Kevin

AU - Levinson, Stephen E

PY - 2005/12/1

Y1 - 2005/12/1

N2 - The term hidden semi-Markov model (HSMM) refers to a large class of stochastic models developed to address some of the shortcomings of hidden Markov models (HMMs). As with HMMs, the underlying sequence of states of a process is modelled as a discrete Markov chain. Unlike HMMs, each state in an HSMM can emit a variable length sequence of observations, with many ways to model duration and observation densities. Parameter estimation in HSMMs is typically done using EM or Viterbi (dynamic programming) algorithms. These algorithms require batch processing of large amounts of data, and so are not useful for online learning. To address this issue, we present here a recursive maximum-likelihood estimation (RMLE) algorithm for online estimation of HSMM parameters, based on a similar method developed for HMMs.

AB - The term hidden semi-Markov model (HSMM) refers to a large class of stochastic models developed to address some of the shortcomings of hidden Markov models (HMMs). As with HMMs, the underlying sequence of states of a process is modelled as a discrete Markov chain. Unlike HMMs, each state in an HSMM can emit a variable length sequence of observations, with many ways to model duration and observation densities. Parameter estimation in HSMMs is typically done using EM or Viterbi (dynamic programming) algorithms. These algorithms require batch processing of large amounts of data, and so are not useful for online learning. To address this issue, we present here a recursive maximum-likelihood estimation (RMLE) algorithm for online estimation of HSMM parameters, based on a similar method developed for HMMs.

UR - http://www.scopus.com/inward/record.url?scp=33749061970&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33749061970&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2005.1532923

DO - 10.1109/MLSP.2005.1532923

M3 - Conference contribution

AN - SCOPUS:33749061970

SN - 0780395174

SN - 9780780395176

T3 - 2005 IEEE Workshop on Machine Learning for Signal Processing

SP - 329

EP - 334

BT - 2005 IEEE Workshop on Machine Learning for Signal Processing

ER -