On adjusted Viterbi training

DOI10.1007/S10440-007-9102-5zbMATH Open1116.62082arXivmath/0406237OpenAlexW2098390086MaRDI QIDQ996733FDOQ996733

Authors: Alexey Koloydenko, Meelis Käärik, J. Lember 

Publication date: 19 July 2007

Published in: Acta Applicandae Mathematicae (Search for Journal in Brave)

Abstract: We study modifications of the Viterbi Training (VT) algorithm to estimate emission parameters in Hidden Markov Models (HMM) in general, and in mixure models in particular. Motivated by applications of VT to HMM that are used in speech recognition, natural language modeling, image analysis, and bioinformatics, we investigate a possibility of alleviating the inconsistency of VT while controlling the amount of extra computations. Specifically, we propose to enable VT to asymptotically fix the true values of the parameters as does the EM algorithm. This relies on infinite Viterbi alignment and an associated with it limiting probability distribution. This paper, however, focuses on mixture models, an important case of HMM, wherein the limiting distribution can always be computed exactly; finding such limiting distribution for general HMM presents a more challenging task under our ongoing investigation. A simulation of a univariate Gaussian mixture shows that our central algorithm (VA1) can dramatically improve accuracy without much cost in computation time. We also present VA2, a more mathematically advanced correction to VT, verify by simulation its fast convergence and high accuracy; its computational feasibility remains to be investigated in future work.

Full work available at URL: https://arxiv.org/abs/math/0406237

Recommendations

• The adjusted Viterbi training for hidden Markov models
• ADJUSTED VITERBI TRAINING
• Extended Viterbi algorithm for hidden Markov process: a transient/steady probabilities approach
• Viterbi training in PRISM
• scientific article; zbMATH DE number 1099037

zbMATH Keywords

EMParameter estimationConsistencyMaximum likelihoodComputational efficiencyHidden Markov modelsBaumConsistentMixture modelsViterbi extractionViterbi training

Mathematics Subject Classification ID

Markov processes: estimation; hidden Markov models (62M05) Inference from stochastic processes (62M99) Non-Markovian processes: estimation (62M09)

Cites Work

• Asymptotic Statistics
• Model-Based Clustering, Discriminant Analysis, and Density Estimation
• Finite mixture models
• A classification EM algorithm for clustering and two stochastic versions
• Biological Sequence Analysis
• Hidden Markov processes
• Inference in hidden Markov models.
• Asymptotic behaviour of classification maximum likelihood estimates
• Title not available (Why is that?)
• Statistical Inference for Probabilistic Functions of Finite State Markov Chains
• Stochastic volatility models as hidden Markov models and statistical applications
• Maximum-likelihood estimation for hidden Markov models
• Consistent and asymptotically normal parameter estimates for hidden Markov models
• Hidden Markov Models for Speech Recognition
• Title not available (Why is that?)
• Multiresolution image classification by hierarchical modeling with two-dimensional hidden Markov models
• Properties of the maximum a posteriori path estimator in hidden Markov models
• Convergence of the maximum a posteriori path estimator in hidden Markov models
• ADJUSTED VITERBI TRAINING
• Title not available (Why is that?)
• A Lagrangian formulation of Zador's entropy-constrained quantization theorem
• Application of the Conditional Population-Mixture Model to Image Segmentation
• The segmental K-means algorithm for estimating parameters of hidden Markov models
• Title not available (Why is that?)
• Global convergence and empirical consistency of the generalized Lloyd algorithm
• A computationally efficient approach to the estimation of two- and three-dimensional hidden Markov models

Cited In (8)

• ADJUSTED VITERBI TRAINING
• MAP segmentation in Bayesian hidden Markov models: a case study
• Handling ties correctly and efficiently in Viterbi training using the Viterbi semiring
• Estimation of Viterbi path in Bayesian hidden Markov models
• Regenerativity of Viterbi process for pairwise Markov models
• Existence of infinite Viterbi path for pairwise Markov models
• Asymptotic risks of Viterbi segmentation
• The adjusted Viterbi training for hidden Markov models

This page was built for publication: On adjusted Viterbi training

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q996733)