Abstract
Hidden Markov models (HMMs) are known for their ability to well model and easily handle variable length time-series. Their use in the case of proportional data modeling has been seldom mentioned in the literature. However, proportional data are a common way of representing large data in a compact fashion and often arise in pattern recognition applications frameworks. HMMs have been first developed for discrete and Gaussian data and their extension to proportional data through the use of Dirichlet distributions is quite recent. The Dirichlet distribution has its limitations and is a special case of the more general generalized Dirichlet (GD) distribution that suffers from less restrictions on the modeled data. We propose here to derive the equations and the methodology of a GD-based HMM and to assess its superiority over a Dirichlet-based HMM (HMMD) through experiments conducted on both synthetic and real data.
Chapter PDF
Similar content being viewed by others
References
Andersson, S., Ryden, T.: Subspace estimation and prediction methods for hidden Markov models. The Annals of Statistics 37(6B), 4131–4152 (2009)
Andrade, E.L., Blunsden, S., Fisher, R.B.: Hidden Markov models for optical flow analysis in crowds. In: ICPR (1), pp. 460–463. IEEE Computer Society (2006)
Bicego, M., Castellani, U., Murino, V.: A hidden Markov model approach for appearance-based 3d object recognition. Pattern Recogn. Lett. 26(16), 2588–2599 (2005)
Blank, M., Gorelick, L., Shechtman, E., Irani, M., Basri, R.: Actions as space-time shapes. In: ICCV, pp. 1395–1402. IEEE Computer Society (2005)
Bouguila, N., Ziou, D.: High-dimensional unsupervised selection and estimation of a finite generalized dirichlet mixture model based on minimum message length. IEEE Transactions on Pattern Analysis and Machine Intelligence 29(10), 1716–1731 (2007)
Chang, W.Y., Gupta, R.D., Richards, D.S.P.: Structural properties of the generalized dirichlet distributions. Contemporary Mathematics 516, 109–124 (2010)
Chaudhry, R., Ravichandran, A., Hager, G.D., Vidal, R.: Histograms of oriented optical flow and Binet-Cauchy kernels on nonlinear dynamical systems for the recognition of human actions. In: CVPR, pp. 1932–1939. IEEE (2009)
Chen, L., Barber, D., Odobez, J.M.: Dynamical dirichlet mixture model. IDIAP-RR 02, IDIAP (2007)
Cholewa, M., Glomb, P.: Estimation of the number of states for gesture recognition with hidden Markov models based on the number of critical points in time sequence. Pattern Recognition Letters 34(5), 574–579 (2013)
Djuric, P.M., Chun, J.H.: An mcmc sampling approach to estimation of nonstationary hidden Markov models. IEEE Transactions on Signal Processing 50(5), 1113–1123 (2002)
Ghahramani, Z., Jordan, M.I.: Factorial hidden Markov models. Machine Learning 29(2-3), 245–273 (1997)
Hjalmarsson, H., Ninness, B.: Fast, non-iterative estimation of hidden Markov models. In: Proc. IEEE Conf. Acoustics, Speech and Signal Process, vol. 4, pp. 2253–2256. IEEE (1998)
Holzmann, H., Munk, A., Suster, M., Zucchini, W.: Hidden Markov models for circular and linear-circular time series. Environmental and Ecological Statistics 13(3), 325–347 (2006)
Jiang, F., Wu, Y., Katsaggelos, A.K.: Abnormal event detection from surveillance video by dynamic hierarchical clustering. In: ICIP (5), pp. 145–148. IEEE (2007)
Maruotti, A.: Mixed hidden Markov models for longitudinal data: An overview. International Statistical Review 79(3), 427–454 (2011)
Maruotti, A., Rocci, R.: A mixed nonhomogeneous hidden Markov model for categorical data, with application to alcohol consumption. Statist. Med. (31), 871–886 (2012)
Rabiner, L.R.: A tutorial on hidden Markov models and selected applications in speech recognition. Proceedings of the IEEE 77(2), 257–286 (1989)
Rabiner, L.R., Juang, B.H.: An introduction to hidden Markov models. IEEE ASSP Magazine 3(1), 4–16 (1986)
Ronning, G.: Maximum-likelihood estimation of dirichlet distribution. Journal of Statistical Computation and Simulation 32, 215–221 (1989)
Wong, T.T.: Parameter estimation for generalized dirichlet distributions from the sample estimates of the first and the second moments of random variables. Computational Statistics and Data Analysis 54(7), 1756–1765 (2010)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Epaillard, E., Bouguila, N. (2014). Hidden Markov Models Based on Generalized Dirichlet Mixtures for Proportional Data Modeling. In: El Gayar, N., Schwenker, F., Suen, C. (eds) Artificial Neural Networks in Pattern Recognition. ANNPR 2014. Lecture Notes in Computer Science(), vol 8774. Springer, Cham. https://doi.org/10.1007/978-3-319-11656-3_7
Download citation
DOI: https://doi.org/10.1007/978-3-319-11656-3_7
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11655-6
Online ISBN: 978-3-319-11656-3
eBook Packages: Computer ScienceComputer Science (R0)