Title
A variational Bayesian methodology for hidden Markov models utilizing Student's-t mixtures
Abstract
The Student's-t hidden Markov model (SHMM) has been recently proposed as a robust to outliers form of conventional continuous density hidden Markov models, trained by means of the expectation-maximization algorithm. In this paper, we derive a tractable variational Bayesian inference algorithm for this model. Our innovative approach provides an efficient and more robust alternative to EM-based methods, tackling their singularity and overfitting proneness, while allowing for the automatic determination of the optimal model size without cross-validation. We highlight the superiority of the proposed model over the competition using synthetic and real data. We also demonstrate the merits of our methodology in applications from diverse research fields, such as human computer interaction, robotics and semantic audio analysis.
Year
DOI
Venue
2011
10.1016/j.patcog.2010.09.001
Pattern Recognition
Keywords
Field
DocType
robust alternative,conventional continuous density,variational bayes,hidden markov model,violence detection,variational bayesian methodology,student s-t mixture,speaker identification,automatic determination,student s-t,markov model,robotic task failure,hidden markov models,em-based method,student's- t distribution,expectation-maximization algorithm,diverse research field,optimal model size,cross validation,pattern recognition,loudspeakers,robots,student s t distribution,human computer interaction,robotics,audio analysis,expectation maximization algorithm,markov processes
Variable-order Bayesian network,Bayesian inference,Pattern recognition,Forward algorithm,Expectation–maximization algorithm,Markov model,Artificial intelligence,Overfitting,Hidden Markov model,Cross-validation,Machine learning,Mathematics
Journal
Volume
Issue
ISSN
44
2
Pattern Recognition
Citations 
PageRank 
References 
21
0.86
20
Authors
2
Name
Order
Citations
PageRank
Sotirios P. Chatzis125024.25
Dimitrios I. Kosmopoulos237827.91