Maximum A Posteriori Approximation of Dirichlet and Beta-Liouville Hidden Markov Models for Proportional Sequential Data Modeling

被引:0
作者
Ali, Samr [1 ]
Bouguila, Nizar [2 ]
机构
[1] Concordia Univ, Dept Elect & Comp Engn, Montreal, PQ, Canada
[2] Concordia Univ, Concordia Inst Informat Syst Engn, Montreal, PQ, Canada
来源
2020 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC) | 2020年
关键词
Hidden Markov models; statistical analysis; Maximum a Posteriori; proportional time series; dynamic textures; infrared action recognition; MIXTURE; RECOGNITION;
D O I
10.1109/smc42975.2020.9283011
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Hidden Markov models (HMM) have recently risen as a key generative machine learning approach for time series data study and analysis. While early works focused only on applying HMMs for speech recognition, HMMs are now prominent in various fields such as stock market forecasting, video classification, and genomics. In this paper, we develop a Maximum A Posteriori (MAP) framework for learning the Dirichlet and Beta-Liouville HMMs that have been proposed recently as an efficient way for modeling sequential proportional data. In contrast to the conventional Baum Welch algorithm, commonly used for learning HMMs, the proposed algorithm places priors for the learning of the desired parameters; hence, regularizing the estimation process. We validate our proposed approach on two challenging real applications; namely, dynamic texture classification and infrared action recognition.
引用
收藏
页码:4081 / 4087
页数:7
相关论文
共 33 条
[1]   Feature Selection for Hidden Markov Models and Hidden Semi-Markov Models [J].
Adams, Stephen ;
Beling, Peter A. ;
Cogill, Randy .
IEEE ACCESS, 2016, 4 :1642-1657
[2]  
Ali S., 2019, 2019 IEEE C COMP VIS
[3]  
Ali S, 2020, IEEE IMAGE PROC, P3279
[4]   Hybrid Generative-Discriminative Generalized Dirichlet-based Hidden Markov Models with Support Vector Machines [J].
Ali, Samr ;
Bouguila, Nizar .
2019 IEEE INTERNATIONAL SYMPOSIUM ON MULTIMEDIA (ISM 2019), 2019, :231-232
[5]   River Runoff Modelling Through Geospatial Techniques-A Case Study of Snow and Glacier Fed Astore River Basin, Northern Pakistan [J].
Ali, Syeda Saleha Fatim ;
Mehmood, Syed Amer ;
Hassan, Mujtaba ;
Latif, Adil .
2019 SIXTH INTERNATIONAL CONFERENCE ON AEROSPACE SCIENCE AND ENGINEERING (ICASE), 2019, :188-199
[6]  
Bouguila N, 2005, IEEE IMAGE PROC, P61
[7]   Online clustering via finite mixtures of Dirichlet and minimum message length [J].
Bouguila, N ;
Ziou, D .
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2006, 19 (04) :371-379
[8]  
Bouguila N, 2005, LECT NOTES ARTIF INT, V3587, P42
[9]  
Bouguila N, 2005, LECT NOTES COMPUT SC, V3686, P172
[10]   Using unsupervised learning of a finite Dirichlet mixture model to improve pattern recognition applications [J].
Bouguila, N ;
Ziou, D .
PATTERN RECOGNITION LETTERS, 2005, 26 (12) :1916-1925