Abstract
Most system identification approaches and statistical inference methods rely on the availability of the analytic knowledge of the probability distribution function of the system output variables. In the case of dynamic systems modelled by hidden Markov chains or stochastic nonlinear state-space models, these distributions as well as that of the state variables themselves, can be unknown or untractable. In that situation, the usual particle Monte Carlo filters for system identification or likelihood-based inference and model selection methods have to rely, whenever possible, on some hazardous approximations and are often at risk. This review shows how a recent nonparametric particle filtering approach can be efficiently used in that context, not only for consistent filtering of these systems but also to restore these statistical inference methods, allowing, for example, consistent particle estimation of Bayes factors or the generalisation of model parameter change detection sequential tests.
Real-life applications of these particle approaches to a microbiological growth model are proposed as illustrations.
Similar content being viewed by others
References
Akashi, H., Kumamoto, H., Nose, K.: Application of Monte Carlo methods to optimal control for linear systems under measurement noise with Markov dependent statistical property. Int. J. Control 22(6), 821–836 (1975)
Bain, A., Crisan, D.: Fundamentals of Stochastic Filtering. Springer, New York (2009)
Baranyi, J., Roberts, T.A.: Mathematics of predictive food microbiology. Int. J. Food Microbiol. 26, 199–218 (1995)
Bartolucci, F., Scaccia, L., Mira, A.: Efficient Bayes factor estimation from the reversible jump output. Biometrika 93(1), 41–52 (2006)
Campillo, F., Rossi, V.: Convolution particle filter for parameter estimation in general state-space models. IEEE Trans. Aerosp. Electron. Syst. 45, 1063–1072 (2009)
Cappé, O., Moulines, E., Rydén, T.: Inference in Hidden Markov Models. Springer, New York (2005)
Chen, G.: Approximate Kalman Filtering. Approximations and Decompositions, vol. 2. World Scientific, Singapore (1993)
Davis, M.: New approach to filtering nonlinear systems. IEEE Proc., Part D 128(5), 166–172 (1981)
Dawid, A.P.: Statistical theory: the prequential approach. J. R. Stat. Soc., A 147, 278–292 (1984)
Dean, T.A., Singh, S.S., Jasra, A., Peters, G.W.: Parameter estimation for hidden Markov models with intractable likelihoods (2011). arXiv:1103.5399v1
Del Moral, P.: Nonlinear filtering using random particles. Theory Probab. Appl. 40(4), 690–701 (1995)
Del Moral, P.: A uniform convergence theorem for the numerical solving of the nonlinear filtering problem. J. Appl. Probab. 35(4), 873–884 (1998)
Del Moral, P., Jacod, J., Interacting Particle Filtering with Discrete Observation. In: Doucet, A., de de Freitas, N., Gordon, N. (eds.), Sequential Monte Carlo Methods in Practice. Statistics for Engineering and Information Science, pp. 43–75. Springer, Berlin (2001)
Del Moral, P., Jacod, J., Protter, P.: The Monte-Carlo method for filtering with discrete-time observations. Probab. Theory Relat. Fields 120, 346–368 (2001)
Del Moral, P., Rigal, G., Salut, G.: Estimation et commande optimale non linéaire: un cadre unifié pour la résolution particulaire. Technical report 2, LAAS/CNRS, contrat DRET-DIGILOG, Toulouse (2011)
Del Moral, P., Doucet, A., Jasra, A.: An adaptive sequential Monte Carlo method for approximate Bayesian computation. Stat. Comput. (2012). doi:10.1007/s11222-011-9271-y
Doucet, A., de Freitas, N., Gordon, N.: Sequential Monte Carlo Methods in Practice. Statistics for Engineering and Information Science. Springer, New York (2001)
Doucet, A., Rousset, M.: In discussion on the paper by Beskos et al. J. R. Stat. Soc. B 68(3), 333–382 (2006)
Ellouze, M., Gauchi, J.P., Augustin, J.C.: Global sensitivity analysis applied to a contamination assessment model of Listeria monocytogenes in cold smoked salmon at consumption. Risk Anal. 30(5), 841–852 (2010)
Fearnhead, P., Papaspiliopoulos, O., Roberts, G.O.: Particle filters for partially observed diffusions. J. R. Stat. Soc. B 70(4), 755–777 (2008)
Gauchi, J.P., Vila, J.P.: Optimal sequential sampling design for improving parametric identification of complex microbiological dynamic systems by nonlinear filtering. In: 7th International Conference on Predictive Modelling in Foods, Dublin, Ireland (2011)
Gauchi, J.P., Bidot, J.C., Augustin, J.C., Vila, J.P.: Identification of complex microbiological dynamic systems by nonlinear filtering. In: 6th ICPMF, Washington (2009)
Gauchi, J.P., Lehuta, S., Mahévas, S.: Constrained Global Sensitivity Analysis: Partial Least Squares Regression Metamodelling and D-optimal Computer Experiment Design. Reliab. Eng. Syst. Saf. (2010, in revision)
Gauchi, J.P., Vila, J.P., Bidot, C., Atlijani, E., Coroller, L., Augustin, J.C., Del Moral, P.: FILTREX: a new software for identification and optimal sampling of experiments for complex microbiological dynamic systems by nonlinear filtering. In: 7th International Conference on Predictive Modelling in Foods, Dublin (2011)
Gordon, N.J., Salmond, D.J., Smith, A.F.M.: Novel approach to nonlinear non-Gaussian Bayesian state estimation. IEE Proc., F 140(2), 107–113 (1993)
Hürzeler, M., Künsch, H.R.: Monte Carlo approximations for general state-space models. J. Comput. Graph. Stat. 7(2), 175–193 (1998)
Jasra, A., Singh, S.S., Martin, J.S., McCoy, E.: Filtering via approximate Bayesian computation. Stat. Comput. (2012). doi:10.1007/s11222-010-9185-0
Jazwinski, A.H.: Stochastic Processes and Filtering Theory. Academic Press, London (1970)
Jeffreys, H.: Theory of Probability. Oxford University Press, Oxford (1961)
Kass, R.E., Raftery, A.E.: Bayes factors. J. Am. Stat. Assoc. 90, 773–795 (1995)
Kitagawa, G.: Non-Gaussian state-space modeling of nonstationary time series. J. Am. Stat. Assoc. 82(400), 1032–1041 (1987)
Kramer, S.C., Sorenson, H.W.: Recursive Bayesian estimation using piece-wise constant approximations. Automatica 21, 789–801 (1988)
Lai, T.L.: Information bounds and quick detection of parameter changes in stochastic systems. IEEE Trans. Inf. Theory 44, 2917–2929 (1998)
LeGLand, F., Oudjane, N.: Stability and uniform approximation of nonlinear filters using the Hilbert metric, and application to particle filters. Ann. Appl. Probab. 14(1), 144–187 (2004)
Liu, J.S.: Monte Carlo Strategies in Scientific Computing. Springer, New York (2001)
Lorden, G.: Procedures for reacting to a change in distribution. Ann. Math. Stat. 42, 1897–1908 (1971)
Magni, L., Raimondo, D.M., Allgöwer, F.: Nonlinear Model Predictive Control. Springer, New York (2009)
Musso, C., Oudjane, N., LeGland, F.: Improving regularised particle filters. In: Doucet, A., de de Freitas, N., Gordon, N. (eds.) Sequential Monte Carlo Methods in Practice. Statistics for Engineering and Information Science, pp. 247–271. Springer, New York (2001)
Oudjane, N.: Stabilité et approximations particulaires en filtrage non linéaires, application au pistage. PhD thesis, Université de Rennes I, France (2000)
Page, E.S.: Continuous inspection schemes. Biometrika 41, 100–115 (1954)
Parzen, E.: On estimation of a probability density function and mode. Ann. Math. Stat. 33, 1065–1076 (1962)
Robert, C., Cornuet, J.M., Marin, J.M., Pillai, N.S.: Lack of confidence in ABC model choice (2011). arXiv:1102.4432
Rodriguez-Fernandez, M., Kucherenko, S., Pantelides, C., Shah, N.: Optimal experimental design based on global sensitivity analysis. In: 17th European Symposium on Computer Aided Process Engineering (2007)
Rossi, V.: Filtrage non linéaire par noyaux de convolution. Application à un procédé de dépollution biologique. Thèse de Docteur en science, ENSAM (2004)
Rossi, V., Vila, J.P.: Approche non paramétrique du filtrage de système non linéaire à temps discret et à paramètres inconnus. C.R. Acad. Sci. Paris, I 340, 759–764 (2005)
Rossi, V., Vila, J.P.: Nonlinear filtering in discrete time: A particle convolution approach. Publ. Inst. Stat. Univ. Paris, L, 3, 71–102 (2006)
Rosso, L.: Modélisation et Microbiologie Prévisionnelle: élaboration D’un Nouvel Outil Pour L’agro-alimentaire. PhD thesis, Université de Lyon I, n o 95-197 (1995)
Šimandl, M., Královec, J.: Filtering, prediction and smoothing with Gaussian sum representation. In: Proceedings of the IFAC 12th Symposium on System Identification, Santa Barbara, USA (2000)
Šimandl, M., Královec, J., Söderström, T.: Advanced point-mass method for nonlinear state estimation. Automatica 42, 1133–1145 (2006)
Sobol, I.: Global sensitivy indices for nonlinear mathematical models and their Monte-Carlo estimates. Math. Comput. Simul. 55, 271–280 (2001)
Sorenson, H.W., Alspach, D.L.: Recursive Bayesian estimation using Gaussian sums. Automatica 7, 465–479 (1971)
Verdier, G., Hilgert, N., Vila, J.P.: Optimality of CUSUM rule approximations in change-point detection problems. Application to nonlinear state-space systems. IEEE Trans. Inf. Theory 54, 5102–5112 (2008)
Vila, J.P.: Nonparametric multi-step prediction in nonlinear state space dynamic systems. Stat. Probab. Lett. 81, 71–76 (2011)
Vila, J.P.: Enhanced consistency of the resampled convolution particle filter. Stat. Probab. Lett. 82, 786–797 (2012)
Vila, J.P., Gauchi, J.P.: Predictive control of stochastic nonlinear state space dynamic systems: a particle nonparametric approach. Technical report, UMR MISTEA (2011)
Vila, J.P., Saley, I.: Estimation de facteurs de Bayes entre modèles dynamiques non linéaires à espace d’état. C.R. Acad. Sci. Paris, Ser. I 347, 429–434 (2009)
Warnes, G.R.: The normal kernel coupler: an adaptave Markov chain Monte Carlo method for efficiently sampling from multi-modal distributions. Technical report 39, Departement of Statistics, University of Washington (2001)
Acknowledgements
The authors would like to thank both reviewers for their careful reading of the manuscript, relevant remarks and new references, which helped to improve the quality of the paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Gauchi, JP., Vila, JP. Nonparametric particle filtering approaches for identification and inference in nonlinear state-space dynamic systems. Stat Comput 23, 523–533 (2013). https://doi.org/10.1007/s11222-012-9327-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-012-9327-7