Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.14279/1711
Title: Hidden Markov models with nonelliptically contoured state densities
Authors: Chatzis, Sotirios P. 
Major Field of Science: Engineering and Technology
Field Category: Computer and Information Sciences
Keywords: Expectation-maximization;Hidden Markov models;Multivariate normal inverse Gaussian (MNIG) distribution;Sequential data modeling
Issue Date: 2010
Source: IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, vol. 32, no. 12, pp. 2297-2304
Volume: 32
Issue: 12
Start page: 2297
End page: 2304
Journal: IEEE Transactions on Pattern Analysis and Machine Intelligence 
Abstract: Hidden Markov models (HMMs) are a popular approach for modeling sequential data comprising continuous attributes. In such applications, the observation emission densities of the HMM hidden states are typically modeled by means of elliptically contoured distributions, usually multivariate Gaussian or Student's-t densities. However, elliptically contoured distributions cannot sufficiently model heavy-tailed or skewed populations which are typical in many fields, such as the financial and the communication signal processing domain. Employing finite mixtures of such elliptically contoured distributions to model the HMM state densities is a common approach for the amelioration of these issues. Nevertheless, the nature of the modeled data often requires postulation of a large number of mixture components for each HMM state, which might have a negative effect on both model efficiency and the training data set's size required to avoid overfitting. To resolve these issues, in this paper, we advocate for the utilization of a nonelliptically contoured distribution, the multivariate normal inverse Gaussian (MNIG) distribution, for modeling the observation densities of HMMs. As we experimentally demonstrate, our selection allows for more effective modeling of skewed and heavy-tailed populations in a simple and computationally efficient manner
URI: https://hdl.handle.net/20.500.14279/1711
ISSN: 19393539
DOI: 10.1109/TPAMI.2010.153
Rights: © IEEE
Type: Article
Affiliation : Imperial College London 
Publication Type: Peer Reviewed
Appears in Collections:Άρθρα/Articles

CORE Recommender
Show full item record

SCOPUSTM   
Citations

18
checked on Nov 9, 2023

WEB OF SCIENCETM
Citations 1

14
Last Week
0
Last month
0
checked on Oct 29, 2023

Page view(s)

489
Last Week
2
Last month
1
checked on Nov 21, 2024

Google ScholarTM

Check

Altmetric


Items in KTISIS are protected by copyright, with all rights reserved, unless otherwise indicated.