Abstract
Hidden Markov models (HMMs) are intensively used in various fields to model and classify data observed along a line (e.g. time). The fit of such models strongly relies on the choice of emission distributions that are most often chosen among some parametric family. In this paper, we prove that finite state space non parametric HMMs are identifiable as soon as the transition matrix of the latent Markov chain has full rank and the emission probability distributions are linearly independent. This general result allows the use of semi- or non-parametric emission distributions. Based on this result we present a series of classification problems that can be tackled out of the strict parametric framework. We derive the corresponding inference algorithms. We also illustrate their use on few biological examples, showing that they may improve the classification performances.




Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.References
Allman, E.S., Matias, C., Rhodes, J.: Identifiability of parameters in latent structure models with many observed variables. Ann. Stat. 37, 3099–3132 (2009)
An, Y., Hu, Y., Hopkins, J., Shum, M.: Identifiability and inference of hidden Markov models. Technical report (2013)
Baudry, J.-P., Raftery, A.E., Celeux, G., Lo, K., Gottardo, R.: Combining mixture components for clustering. J. Comput. Gr. Stat. 19(2), 332–353 (2010)
Benaglia, T., Chauveau, D., Hunter, D.R.: An EM-like algorithm for semi-and nonparametric estimation in multivariate mixtures. J. Comput. Gr. Stat. 18(2), 505–526 (2009)
Bérard, C., Martin-Magniette, M.L., Brunaud, V., Aubourg, S., Robin, S.: Unsupervised classification for tiling arrays: ChIP-chip and transcriptome. Stat. Appl. Genet. Mol. Biol. 10(1), 1–22 (2011)
Bordes, L., Mottelet, S., Vandekerkhove, P.: Semiparametric estimation of a two components mixture model. Ann. Stat. 34, 1204–1232 (2006a)
Bordes, L., Delmas, C., Vandekerkhove, P.: Semiparametric estimation of a two-component mixture model where one component is known. Scand. J. Stat. 33(4), 733–752 (2006b)
Butucea, C., Vandekerkhove, P.: Semiparametric mixtures of symmetric distributions. Scand. J. Stat. 41(1), 227–239 (2014)
Cappé, O., Moulines, E., Rydén, T.: Inference Hidden Markov Models. Springer, New York (2005)
Couvreur, L., Couvreur, C.: Wavelet based non-parametric HMMs: theory and methods. In: ICASSP ’00 Proceedings, pp. 604–607 (2000)
Dempster, A.P., Laird, N.M., Rubin, D.B.: Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B 39, 1–38 (1977)
DeSantis, S.M., Bandyopadhyay, D.: Hidden Markov models for zero-inflated Poisson counts with an application to substance use. Stat. Med. 30(14), 1678–1694 (2011)
Donoho, D.L., Johnstone, I.M., Kerkyacharian, G., Picard, D.: Density estimation by wavelet thresholding. Ann. Stat. 24(2), 508–539 (1996)
Du, J., Rozowsky, J.S., Korbel, J.O., Zhang, Z.D., Royce, T.E., Schultz, M.H., Snyder, M., Gerstein, M.: A supervised hidden Markov model framework for efficiently segmenting tiling array data in transcriptional and chIP-chip experiments: systematically incorporating validated biological knowledge. Bioinformatics 22(24), 3016–3024 (2006)
Dumont, T., Le Corff, S.: Nonparametric regression on hidden phi-mixing variables: identifiability and consistency of a pseudo-likelihood based estimation procedure. Technical report, arXiv:1209.0633D (Sep., 2012)
Durot, C., Huet, S., Koladjo, F., Robin, S.: Least-squares estimation of a convex discrete distribution. Comput. Stat. Data Anal. 67, 282–298 (2013)
Gassiat, E., Rousseau, J.: Non parametric finite translation hidden Markov models and extensions. Bernoulli. to appear (2014)
Hall, P., Zhou, X.-H.: Nonparametric estimation of component distributions in a multivariate mixture. Ann. Stat. 31(1), 201–224 (2003)
Hsu, D., Kakade, S.M., Zhang, T.: A spectral algorithm for learning hidden Markov models. J. Comput. Syst. Sci. 78, 1460–1480 (2012)
Jin, N., Mokhtarian, F.: A non-parametric HMM learning method for shape dynamics with application to human motion recognition. In: 18th International Conference on Pattern Recognition, 2006. ICPR 2006, vol. 2, pp. 29–32. IEEE (2006)
Lambert, M., Whiting, J., Metcalfe, A.: A non-parametric hidden Markov model for climate state identification. Hydrol. Earth Syst. Sci. 7(5), 652–667 (2003)
Lefèvre, F.: Non-parametric probability estimation for HMM-based automatic speech recognition. Comput. Speach Lang. 17, 113–136 (2003)
Levine, M., Hunter, D.R., Chauveau, D.: Maximum smoothed likelihood for multivariate mixtures. Biometrika. 98(2), 403–416 (2011)
Li, H., Zhang, K., Jiang, T.: The regularized EM algorithm. In: Proceedings of the National Conference on Artificial Intelligence, vol. 20, p. 807. Menlo Park, CA; Cambridge, MA; London; AAAI Press; MIT Press; 1999 (2005)
Lin, T.I., Lee, J.C., Yen, S.Y.: Finite mixture modelling using the skew normal distribution. Stat. Sin. 17(3), 909–927 (2007)
Massart, P.: Concentration inequalities and model selection. Volume 1896 of Lecture Notes in Mathematics. Lectures from the 33rd Summer School on Probability Theory held in Saint-Flour, July 6–23, 2003, With a foreword by Jean Picard. Springer, Berlin (2007)
Maugis, C., Michel, B.: A non asymptotic penalized criterion for Gaussian mixture model selection. ESAIM Probab. Stat. 15, 41–68 (2011)
Olteanu, M., Ridgway, J., et al.: Hidden Markov models for time series of counts with excess zeros. Proc. ESANN 2012, 133–138 (2012)
Petrie, T.: Probabilistic functions of finite state Markov chains. Ann. Math. Stat. 40, 97–115 (1969)
Shang, L., Chan, K.: Nonparametric discriminant HMM and application to facial expression recognition. In: 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pp. 2090–2096 (2009)
Titterington, D.M., Smith, A.F.M., Makov, U.E.: Statistical Analysis of Finite Mixture Distributions. Wiley Series in Probability and Mathematical Statistics. Applied Probability and Statistics. Wiley, Chichester (1985)
Tune, P., Nguyen, H. X., Roughan, M.: Hidden Markov model identifiability via tensors. In: 2013 IEEE International Symposium on Information Theory Proceedings (ISIT), pp. 2299–2303. IEEE (2013)
van de Geer, S.A.: Empirical processes in M-estimation. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge (2000)
Vernet, E.: Posterior consistency for nonparametric Hidden Markov Models with finite state space. Technical report, arXiv:1311.3092V (2013)
Volant, S., Bérard, C., Martin-Magniette, M.-L., Robin, S.: Hidden Markov models with mixtures as emission distributions. Stat. Comput. 1–12 (2013). doi:10.1007/s11222-013-9383-7
Yakowitz, S.J., Spragins, J.D.: On the identifiability of finite mixtures. Ann. Math. Stat. 39, 209–214 (1968)
Zhai, Z., Ku, S.-Y., Luan, Y., Reinert, G., Waterman, M.S., Sun, F.: The power of detecting enriched patterns: an HMM approach. J. Comput. Biol. 17(4), 581–592 (2010)
Acknowledgments
The authors want to thank Caroline Bérard for providing the transcriptomic tiling array data. Part of this work was supported by the ABS4NGS ANR project (ANR-11-BINF-0001-06).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Gassiat, E., Cleynen, A. & Robin, S. Inference in finite state space non parametric Hidden Markov Models and applications. Stat Comput 26, 61–71 (2016). https://doi.org/10.1007/s11222-014-9523-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-014-9523-8