Statistical Science

A Comparison of Inferential Methods for Highly Nonlinear State Space Models in Ecology and Epidemiology

Matteo Fasiolo, Natalya Pya, and Simon N. Wood

Full-text: Access denied (no subscription detected)

We're sorry, but we are unable to provide you with the full text of this article because we are not able to identify you as a subscriber. If you have a personal subscription to this journal, then please login. If you are already logged in, then you may need to update your profile to register your subscription. Read more about accessing full-text

Abstract

Highly nonlinear, chaotic or near chaotic, dynamic models are important in fields such as ecology and epidemiology: for example, pest species and diseases often display highly nonlinear dynamics. However, such models are problematic from the point of view of statistical inference. The defining feature of chaotic and near chaotic systems is extreme sensitivity to small changes in system states and parameters, and this can interfere with inference. There are two main classes of methods for circumventing these difficulties: information reduction approaches, such as Approximate Bayesian Computation or Synthetic Likelihood, and state space methods, such as Particle Markov chain Monte Carlo, Iterated Filtering or Parameter Cascading. The purpose of this article is to compare the methods in order to reach conclusions about how to approach inference with such models in practice. We show that neither class of methods is universally superior to the other. We show that state space methods can suffer multimodality problems in settings with low process noise or model misspecification, leading to bias toward stable dynamics and high process noise. Information reduction methods avoid this problem, but, under the correct model and with sufficient process noise, state space methods lead to substantially sharper inference than information reduction methods. More practically, there are also differences in the tuning requirements of different methods. Our overall conclusion is that model development and checking should probably be performed using an information reduction method with low tuning requirements, while for final inference it is likely to be better to switch to a state space method, checking results against the information reduction approach.

Article information

Source
Statist. Sci., Volume 31, Number 1 (2016), 96-118.

Dates
First available in Project Euclid: 10 February 2016

Permanent link to this document
https://projecteuclid.org/euclid.ss/1455115916

Digital Object Identifier
doi:10.1214/15-STS534

Mathematical Reviews number (MathSciNet)
MR3458595

Zentralblatt MATH identifier
06946214

Keywords
Nonlinear dynamics state space models particle filters approximate Bayesian computation statistical ecology

Citation

Fasiolo, Matteo; Pya, Natalya; Wood, Simon N. A Comparison of Inferential Methods for Highly Nonlinear State Space Models in Ecology and Epidemiology. Statist. Sci. 31 (2016), no. 1, 96--118. doi:10.1214/15-STS534. https://projecteuclid.org/euclid.ss/1455115916


Export citation

References

  • Anderson, C. N. K., Hsieh, C.-h., Sandin, S. A., Hewitt, R., Hollowed, A., Beddington, J., May, R. M. and Sugihara, G. (2008). Why fishing magnifies fluctuations in fish abundance. Nature 452 835–839.
  • Andrieu, C. and Doucet, A. (2003). Online expectation-maximization type algorithms for parameter estimation in general state space models. In IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP ’03) 6 69–72. IEEE, New York.
  • Andrieu, C., Doucet, A. and Holenstein, R. (2010). Particle Markov chain Monte Carlo methods. J. R. Stat. Soc. Ser. B. Stat. Methodol. 72 269–342.
  • Andrieu, C., Doucet, A. and Tadic, V. B. (2005). On-line parameter estimation in general state–space models. In Decision and Control, 2005 and 2005 European Control Conference. CDC-ECC’05. 44th IEEE Conference on 332–337. IEEE, Piscataway, NJ.
  • Andrieu, C. and Roberts, G. O. (2009). The pseudo-marginal approach for efficient Monte Carlo computations. Ann. Statist. 37 697–725.
  • Beaumont, M. A., Zhang, W. and Balding, D. J. (2002). Approximate Bayesian computation in population genetics. Genetics 162 2025–2035.
  • Berliner, L. M. (1992). Statistics, probability and chaos. Statist. Sci. 7 69–122.
  • Bhadra, A., Ionides, E. L., Laneri, K., Pascual, M., Bouma, M. and Dhiman, R. C. (2011). Malaria in Northwest India: Data analysis via partially observed stochastic differential equation models driven by Lévy noise. J. Amer. Statist. Assoc. 106 440–451.
  • Blum, M. G. B. (2010). Approximate Bayesian computation: A nonparametric perspective. J. Amer. Statist. Assoc. 105 1178–1187.
  • Blum, M. G. B., Nunes, M. A., Prangle, D. and Sisson, S. A. (2013). A comparative review of dimension reduction methods in approximate Bayesian computation. Statist. Sci. 28 189–208.
  • Carlin, B. P., Polson, N. G. and Stoffer, D. S. (1992). A Monte Carlo approach to nonnormal and nonlinear state–space modeling. J. Amer. Statist. Assoc. 87 493–500.
  • Cash, R. A., Music, S. I., Libonati, J. P., Craig, J. P., Pierce, N. F. and Hornick, R. B. (1974). Response of man to infection with Vibrio cholerae. II. Protection from illness afforded by previous disease and vaccine. J. Infectous Diseases 130 325–333.
  • Chan, K.-S. and Tong, H. (2001). Chaos: A Statistical Perspective. Springer, New York.
  • Dean, T. A., Singh, S. S., Jasra, A. and Peters, G. W. (2011). Parameter estimation for hidden Markov models with intractable likelihoods. Preprint. Available at arXiv:1103.5399.
  • Desharnais, R. A., Costantino, R. F., Cushing, J. M., Henson, S. M. and Dennis, B. (2001). Chaos and population control of insect outbreaks. Ecology Letters 4 229–235.
  • Doucet, A., Godsill, S. and Andrieu, C. (2000). On sequential Monte Carlo sampling methods for Bayesian filtering. Stat. Comput. 10 197–208.
  • Doucet, A. and Johansen, A. M. (2009). A tutorial on particle filtering and smoothing: Fifteen years later. In The Oxford Handbook of Nonlinear Filtering 656–704. Oxford Univ. Press, Oxford.
  • Doucet, A., Pitt, M., Deligiannidis, G. and Kohn, R. (2012). Efficient implementation of Markov chain Monte Carlo when using an unbiased likelihood estimator. Preprint. Available at arXiv:1210.1871.
  • Drovandi, C. C., Pettitt, A. N. and Lee, A. (2015). Bayesian indirect inference using a parametric auxiliary model. Statist. Sci. 30 72–95.
  • Earn, D. J., Rohani, P. and Grenfell, B. T. (1998). Persistence, chaos and synchrony in ecology and epidemiology. Proc. R. Soc. Lond. Ser. B: Biol. Sci. 265 7–10.
  • Fasiolo, M., Pya, N. and Wood, S. N. (2016). Supplement to “A Comparison of Inferential Methods for Highly Nonlinear State Space Models in Ecology and Epidemiology.” DOI:10.1214/15-STS534SUPP.
  • Fearnhead, P. and Prangle, D. (2012). Constructing summary statistics for approximate Bayesian computation: Semi-automatic approximate Bayesian computation. J. R. Stat. Soc. Ser. B. Stat. Methodol. 74 419–474.
  • Geweke, J. and Tanizaki, H. (2001). Bayesian estimation of state–space models using the Metropolis–Hastings algorithm within Gibbs sampling. Comput. Statist. Data Anal. 37 151–170.
  • Glass, R. I., Becker, S., Huq, M. I., Stoll, B. J., Khan, M., Merson, M. H., Lee, J. V. and Black, R. E. (1982). Endemic cholera in rural Bangladesh, 1966–1980. American. J. Epidemiology. 116 959–970.
  • Gordon, N. J., Salmond, D. J. and Smith, A. F. (1993). Novel approach to nonlinear/non-Gaussian Bayesian state estimation. In IEE Proc. F (Radar and Signal Process.) 140 107–113. IET, Stevenage.
  • Grenfell, B. T. (1992). Chance and chaos in measles dynamics. J. R. Stat. Soc. Ser. B. Stat. Methodol. 383–398.
  • Grenfell, B. T., Bjørnstad, O. N. and Finkenstädt, B. F. (2002). Dynamics of measles epidemics: Scaling noise, determinism, and predictability with the TSIR model. Ecological Monograph 72 185–202.
  • Grenfell, B. T., Kleczkowski, A., Gilligan, C. A. and Bolker, B. M. (1995). Spatial heterogeneity, nonlinear dynamics and chaos in infectious diseases. Stat. Methods Med. Res. 4 160–183.
  • Gurney, W. S. C., Blythe, S. P. and Nisbet, R. M. (1980). Nicholson’s blowflies revisited. Nature 287 17–21.
  • He, D., Ionides, E. L. and King, A. A. (2010). Plug-and-play inference for disease dynamics: Measles in large and small populations as a case study. Journal of the Royal Society Interface 7 271–283.
  • Ionides, E. L., Bretó, C. and King, A. A. (2006). Inference for nonlinear dynamical systems. Proc. Natl. Acad. Sci. USA 103 18438–18443.
  • Ionides, E. L., Bhadra, A., Atchadé, Y. and King, A. (2011). Iterated filtering. Ann. Statist. 39 1776–1802.
  • Kantas, N., Doucet, A., Singh, S. S., Maciejowski, J. M. and Chopin, N. (2014). On particle methods for parameter estimation in state–space models. Preprint. Available at arXiv:1412.8695.
  • Kausrud, K. L., Mysterud, A., Steen, H., Vik, J. O., Østbye, E., Cazelles, B., Framstad, E., Eikeset, A. M., Mysterud, I., Solhøy, T. et al. (2008). Linking climate change to lemming cycles. Nature 456 93–97.
  • Kendall, B. E., Ellner, S. P., McCauley, E., Wood, S. N., Briggs, C. J., Murdoch, W. W. and Turchin, P. (2005). Population cycles in the pine looper moth: Dynamical tests of mechanistic hypotheses. Ecological Monograph 75 259–276.
  • King, A. A., Ionides, E. L., Pascual, M. and Bouma, M. J. (2008). Inapparent infections and cholera dynamics. Nature 454 877–880.
  • King, A. A., Nguyen, D. andIonides, E. L. (2015). Statistical Inference for Partially Observed Markov Processes via the R Package pomp. Preprint. Available at arXiv:1509.00503.
  • Kitagawa, G. (1998). A self-organising state–space model. J. Amer. Statist. Assoc. 93 1203–1215.
  • Klaas, M., De Freitas, N. and Doucet, A. (2012). Toward practical N2 Monte Carlo: The marginal particle filter. Preprint. Available at arXiv:1207.1396.
  • Koelle, K., Rodó, X., Pascual, M., Yunus, M. and Mostafa, G. (2005). Refractory periods and climate forcing in cholera dynamics. Nature 436 696–700.
  • Lavine, J. S., King, A. A., Andreasen, V. and Bjørnstad, O. N. (2013). Immune boosting explains regime-shifts in prevaccine-era pertussis dynamics. PloS One 8 e72086.
  • Liu, J. and West, M. (2001). Combined parameter and state estimation in simulation-based filtering. In Sequential Monte Carlo Methods in Practice 197–223. Springer, New York.
  • Malik, S. and Pitt, M. K. (2011). Particle filters for continuous likelihood evaluation and maximisation. J. Econometrics 165 190–209.
  • Marjoram, P., Molitor, J., Plagnol, V. and Tavare, S. (2003). Markov chain Monte Carlo without likelihoods. Proc. Natl. Acad. Sci. USA 100 15324–15328.
  • Meeds, E. and Welling, M. (2014). GPS-ABC: Gaussian process surrogate approximate Bayesian computation. Preprint. Available at arXiv:1401.2838.
  • Nemeth, C., Fearnhead, P. and Mihaylova, L. (2013). Particle approximations of the score and observed information matrix for parameter estimation in state space models with linear computational cost. Preprint. Available at arXiv:1306.0735.
  • Nicholson, A. J. (1954). An outline of the dynamics of animal populations. Aust. J. Zoology 2 9–65.
  • Nicholson, A. J. (1957). The self-adjustment of populations to change. In Cold Spring Harbor Symposia on Quantitative Biology 22 153–173. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.
  • Niemi, J. and West, M. (2010). Adaptive mixture modeling Metropolis methods for Bayesian analysis of nonlinear state–space models. J. Comput. Graph. Statist. 19 260–280.
  • Nunes, M. A. and Balding, D. J. (2010). On optimal selection of summary statistics for approximate Bayesian computation. Stat. Appl. Genet. Mol. Biol. 9 Art. 34, 16.
  • Owen, J., Wilkinson, D. J. and Gillespie, C. S. (2014). Likelihood free inference for Markov processes: A comparison. Preprint. Available at arXiv:1410.0524.
  • Pitt, M. K. and Shephard, N. (1999). Filtering via simulation: Auxiliary particle filters. J. Amer. Statist. Assoc. 94 590–599.
  • Polson, N. G., Stroud, J. R. and Müller, P. (2008). Practical filtering with sequential parameter learning. J. R. Stat. Soc. Ser. B. Stat. Methodol. 70 413–428.
  • Poyiadjis, G., Doucet, A. and Singh, S. S. (2011). Particle approximations of the score and observed information matrix in state space models with application to parameter estimation. Biometrika 98 65–80.
  • Ramsay, J. O., Hooker, G., Campbell, D. and Cao, J. (2007). Parameter estimation for differential equations: A generalized smoothing approach. J. R. Stat. Soc. Ser. B. Stat. Methodol. 69 741–796.
  • Roberts, G. O. and Stramer, O. (2001). On inference for partially observed nonlinear diffusion models using the Metropolis–Hastings algorithm. Biometrika 88 603–621.
  • Sherlock, C., Thiery, A. H., Roberts, G. O. and Rosenthal, J. S. (2015). On the efficiency of pseudo-marginal random walk Metropolis algorithms. Ann. Statist. 43 238–275.
  • Snyder, C., Bengtsson, T., Bickel, P. and Anderson, J. (2008). Obstacles to high-dimensional particle filtering. Monthly Weather Review 136 4629–4640.
  • Toni, T., Welch, D., Strelkowa, N., Ipsen, A. and Stumpf, M. P. (2009). Approximate Bayesian computation scheme for parameter inference and model selection in dynamical systems. J. R. Soc. Inter. 6 187–202.
  • Turchin, P. and Ellner, S. P. (2000). Living on the edge of chaos: Population dynamics of Fennoscandian voles. Ecology 81 3099–3116.
  • Wood, S. N. (2010). Statistical inference for noisy nonlinear ecological dynamic systems. Nature 466 1102–1104.
  • Yang, G.-J., Bradshaw, C. J., Whelan, P. I. and Brook, B. W. (2008). Importance of endogenous feedback controlling the long-term abundance of tropical mosquito species. Population Ecology 50 293–305.

Supplemental materials

  • Supplement to “A Comparison of Inferential Methods for Highly Nonlinear State Space Models in Ecology and Epidemiology”. The supplement describes how the likelihood of a discrete SSM can be computed exactly and it contains additional details regarding the examples considered in Section 5.