Skip to Main content Skip to Navigation
Conference papers

Assessment of various initialization strategies for the Expectation-Maximization algorithm for Hidden Semi-Markov Models with multiple categorical sequences

Brice Olivier 1 Anne Guérin-Dugué 2 Jean-Baptiste Durand 1
1 MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems
Inria Grenoble - Rhône-Alpes, Grenoble INP - Institut polytechnique de Grenoble - Grenoble Institute of Technology, LJK - Laboratoire Jean Kuntzmann
2 CLIPS - IMAG - Communication Langagière et Interaction Personne-Système
Abstract : In this study, we propose a method called sequence breaking framework to search high local maximum of the likelihood by providing starting values based on the observations for the Expectation-Maximization algorithm, for Hidden semi-Markov model parameter estimation. The method is shown to be efficient on several datasets with multiple categorical sequences.
Keywords : multiple categorical sequences Hidden semi-Markov models choosing starting values for the EM algorithm
Document type :
Conference papers
Complete list of metadata

Cited literature [12 references]  Display  Hide  Download
https://hal.inria.fr/hal-02129122
Contributor : Brice Olivier Connect in order to contact the contributor
Submitted on : Tuesday, May 14, 2019 - 3:55:21 PM
Last modification on : Thursday, January 20, 2022 - 5:26:16 PM

File

Vignette du document
olivier2019assessment.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02129122, version 1

Collections

IMAG | CNRS | INRIA | LJK | LJK_PS | LJK_PS_MISTIS | UGA

Citation

Brice Olivier, Anne Guérin-Dugué, Jean-Baptiste Durand. Assessment of various initialization strategies for the Expectation-Maximization algorithm for Hidden Semi-Markov Models with multiple categorical sequences. JdS 2019 - 51èmes Journées de Statistique, Jun 2019, Vandœuvre-lès-Nancy, France. pp.1-7. ⟨hal-02129122⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Les métriques sont temporairement indisponibles

 

Contact                    Legal Notice                    Personal Data