Abstract
We consider Bayesian parameter inference associated to partially-observed stochastic processes that start from a set B 0 and are stopped or killed at the first hitting time of a known set A. Such processes occur naturally within the context of a wide variety of applications. The associated posterior distributions are highly complex and posterior parameter inference requires the use of advanced Markov chain Monte Carlo (MCMC) techniques. Our approach uses a recently introduced simulation methodology, particle Markov chain Monte Carlo (PMCMC) (Andrieu et al. 2010), where sequential Monte Carlo (SMC) (Doucet et al. 2001; Liu 2001) approximations are embedded within MCMC. However, when the parameter of interest is fixed, standard SMC algorithms are not always appropriate for many stopped processes. In Chen et al. (2005), Del Moral (2004), the authors introduce SMC approximations of multi-level Feynman-Kac formulae, which can lead to more efficient algorithms. This is achieved by devising a sequence of sets from B 0 to A and then performing the resampling step only when the samples of the process reach intermediate sets in the sequence. The choice of the intermediate sets is critical to the performance of such a scheme. In this paper, we demonstrate that multi-level SMC algorithms can be used as a proposal in PMCMC. In addition, we introduce a flexible strategy that adapts the sets for different parameter proposals. Our methodology is illustrated on the coalescent model with migration.
Similar content being viewed by others
References
Andrieu, C., Doucet, A., Holenstein, R.: Particle Markov chain Monte Carlo methods (with discussion). J. R. Stat. Soc. B 72, 269–342 (2010)
Andrieu, C., Doucet, A., Tadic, V.: On-line parameter estimation in general state-space models using pseudo-likelihood. Technical report, University of Bristol (2009)
Andrieu, C., Roberts, G.O.: The pseudo-marginal approach for efficient Monte Carlo computations. Ann. Stat. 37, 697–725 (2009)
Bahlo, M., Griffiths, R.C.: Coalescence time for two genes from a subdivided population. J. Math. Biol. 43, 397–410 (2000)
Bibbona, E., Ditlevsen, S.: Estimation in discretely observed Markov processes killed at a threshold. Scand. J. Stat. (2012, to appear). arXiv:1011.1356v1
Blom, H.A.P., Bakker, G.J., Krystul, J.: Probabilistic reachability analysis for large scale stochastic hybrid systems. In: Proc. 46th IEEE Conf. Dec. Contr., New Orleans, USA (2007)
Casella, B., Roberts, G.O.: Exact Monte Carlo simulation of killed diffusions. Adv. Appl. Probab. 40, 273–291 (2008)
Cerou, F., Guyader, A.: Adaptive multilevel splitting for rare-events analysis. Stoch. Anal. Appl. 25, 417–433 (2007)
Cerou, F., Del Moral, P., Furon, T., Guyader, A.: Sequential Monte Carlo for rare event estimation. Stat. Comput. 22, 795–808 (2012)
Cerou, F., Del Moral, P., Guyader, A.: A non asymptotic variance theorem for unnormalized Feynman-Kac particle models. Ann. Inst. Henri Poincaré Probab. Stat. 47, 629–649 (2011)
Chen, Y., Xie, J., Liu, J.S.: Stopping-time resampling for sequential Monte Carlo methods. J. R. Stat. Soc. B 67, 199–217 (2005)
Chopin, N., Jacob, P., Papaspiliopoulos, O.: SMC2: A sequential Monte Carlo algorithm with particle Markov chain Monte Carlo updates. J. R. Stat. Soc. B (2012, to appear) arXiv:1101.1528v2
Dean, T., Dupuis, P.: Splitting for rare event simulation: a large deviations approach to design and analysis. Stoch. Process. Appl. 119, 562–587 (2009)
De Iorio, M., Griffiths, R.C.: Importance sampling on coalescent histories. II: Subdivided population models. Adv. Appl. Probab. 36, 434–454 (2004)
Del Moral, P.: Feynman-Kac Formulae: Genealogical and Interacting Particle Systems with Applications. Springer, New York (2004)
Del Moral, P., Garnier, J.: Genealogical particle analysis of rare events. Ann. Appl. Probab. 15, 2496–2534 (2005)
Del Moral, P., Doucet, A., Jasra, A.: Sequential Monte Carlo samplers. J. R. Stat. Soc. B 68, 411–436 (2006)
Doucet, A., De Freitas, J.F.G., Gordon, N.J.: Sequential Monte Carlo Methods in Practice. Springer, New York (2001)
Glasserman, P., Heidelberger, P., Shahabuddin, P., Zajic, T.: Multi-level splitting for estimating rare event probabilities. Oper. Res. 47, 585–600 (1999)
Gorur, D., Teh, Y.W.: An efficient sequential Monte Carlo algorithm for coalescent clustering. Adv. Neural Inf. Process. Syst. 521–528 (2008)
Griffiths, R.C., Tavare, S.: Simulating probability distributions in the coalescent. Theor. Popul. Biol. 46, 131–159 (1994)
Johansen, A.M., Del Moral, P., Doucet, A.: Sequential Monte Carlo samplers for rare events. In: Proc. 6th Interl. Works. Rare Event Simul., Bamberg, Germany (2006)
Kantas, N., Doucet, A., Singh, S.S., Maciejowski, J.M., Chopin, N.: On particle methods for parameter estimation in general state-space models. Technical report, Imperial College London (2011)
Kingman, J.F.C.: On the genealogy of large populations. J. Appl. Probab. 19, 27–43 (1982)
Lee, A., Yau, C., Giles, M., Doucet, A., Holmes, C.C.: On the utility of graphics cards to perform massively parallel implementation of advanced Monte Carlo methods. J. Comput. Graph. Stat. 19, 769–789 (2010)
Lezaud, P., Krystul, J., Le Gland, F.: Sampling per mode simulation for switching diffusions. In: Proc. 8th Internl. Works. Rare-Event Simul.. RESIM, Cambridge (2010)
Liu, J.S.: Monte Carlo Strategies in Scientific Computing. Springer, New York (2001)
Roberts, G.O., Rosenthal, J.S.: Coupling and ergodicity of adaptive MCMC. J. Appl. Probab. 44, 458–475 (2007)
Roberts, G.O., Rosenthal, J.S.: Quantitative non-geometric convergence bounds for independence samplers. Methodol. Comput. Appl. Probab. 13, 391–403 (2011)
Sadowsky, J.S., Bucklew, J.A.: On large deviation theory and asymptotically efficient Monte Carlo estimation. IEEE Trans. Inf. Theory 36, 579–588 (1990)
Stephens, M., Donelly, P.: Inference in molecular population genetics (with discussion). J. R. Stat. Soc. B 62, 605–655 (2000)
Whiteley, N.: Discussion of particle Markov chain Monte Carlo methods. J. R. Stat. Soc. B 72, 306–307 (2010)
Acknowledgements
We thank Arnaud Doucet and Maria De Iorio for many conversations on this work. The first author was supported by a Ministry of Education grant. The second author was kindly supported by the EPSRC programme grant on Control For Energy and Sustainability EP/G066477/1. We thank two referees and an associate editor for comments that vastly improved the paper.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Proof of Proposition 4.1
The result is a straight forward application of Theorem 6 of Roberts and Rosenthal (2011), which adapted to our notation states:
where the conditional expectation is the expectation w.r.t. the SMC algorithm (i.e. Ξ∼ψ θ ) and the outer expectation is w.r.t. the PIMH target (i.e. \(\xi\sim\pi_{p}^{N}\)). We also denote the estimate of the normalising constant as \(\hat{Z}_{p}(\cdot )\) with ⋅ denoting which random variables generate the estimate.
Now, clearly via (A1),
with the convention that τ 0=0. Thus, it follows that
and we obtain:
Note that by assumption, \(Z_{p}(\varXi) (\rho\varphi )^{2\sum_{n=1}^{p}\bar{\tau}_{n}}\leq1\), and thus we have
Given Del Moral (2004, Theorem 7.4.2, Eq. (7.17), p. 239) and the fact that γ θ is defined to be strictly positive in (A1), we have that the SMC approximation \(\hat {Z}_{p}(\cdot)\) is an unbiased estimate of the normalising constant Z p :
and we can easily conclude. □
Proof of Proposition 4.2
The proof of parts 1 and 2 follows the line of arguments used in Theorem 4 of Andrieu et al. (2010), which we will adapt to our set-up. The main difference lies in the multi-level construction and second statement regarding the marginal of \(\bar{\pi}^{N}\). For the validity of the multi-level set-up, we will rely on Proposition 3.1.
Suppose we design a Metropolis-Hastings kernel with invariant density \(\bar{\pi}^{N}\) and use a proposal \(q^{N}(\theta,k,\bar {\mathcal{X}}_{1:p}, \bar{\mathbf{a}}_{1:p-1})=\psi_{\theta}(\bar {\mathcal{X}}_{1:p},\bar{\mathbf{a}}_{1:p-1})f(k\mid W_{p})\bar {q}(\theta(i-1)\mid \theta^{\prime})=\psi_{\theta}(\bar{\mathcal {X}}_{1:p},\bar{\mathbf{a}}_{1:p-1})\bar{W}_{p}^{(k)}\bar{q}(\theta \mid \theta^{\prime})\). Then
where we denote the normalising constant of the posterior in (1) as:
Therefore, the Metropolis-Hastings procedure to sample from \(\bar{\pi }_{p}^{N}\) will be as in Algorithm 4.
Alternatively using similar arguments, one may write
Summing over k and using the unbiased property of the SMC algorithm in Eq. (13), it follows that \(\bar{\pi }_{p}^{N}(\cdot)\) admits \(\bar{\pi}(\theta)\) as a marginal, so the proof of part 1 is complete.
Part 2 is a direct consequence of Theorem 1 in Andrieu and Roberts (2009) and Assumption (A2). □
Proof of Proposition 4.3
The proof is similar to that of Proposition 4.2. For the proof of the first statement of part 1, one repeats the same arguments as for Proposition 4.2, with the difference being in the inclusion of \(\bar{\varLambda}_{\theta}(v)\) for \(\bar {\pi}^{N}\) and \(\bar{q}^{N}\). For the second statement, to get the marginal of \(\bar{\pi}^{N}\), one re-writes the target as
Let \(\bar{\pi}_{p}^{N}(\theta)\) denote the marginal of \(\bar {\pi}_{p}^{N}(\cdot)\) obtained in Proposition 4.2. Using (11) and the conditional independence of v and \(\bar{\mathcal{X}}_{1:p(v)},\bar{\mathbf {a}}_{1:p(v)-1}\), then for the marginal of \(\bar{\pi}^{N}(\cdot)\) w.r.t. v, \(\bar{\mathcal{X}}_{1:p(v)},\bar{\mathbf {a}}_{1:p(v)-1}\), k we have that
where the summing over k and integrating w.r.t. \(\bar{\mathcal {X}}_{1:p(v)}, \bar{\mathbf{a}}_{1:p(v)-1}\) is as in Proposition 4.2.
For part 2 note that the conditional density given k and v and θ of \(\mathcal{X}_{1:p(v)}^{(k)}\) is
Hence the sequence \((\theta(i),\mathcal {X}_{1:p(v)}^{(k)}(i) )_{i\geq0}\) satisfies the required property as a direct consequence of Theorem 1 in Andrieu and Roberts (2009) and Assumption (A2). □
Rights and permissions
About this article
Cite this article
Jasra, A., Kantas, N. & Persing, A. Bayesian parameter inference for partially observed stopped processes. Stat Comput 24, 1–20 (2014). https://doi.org/10.1007/s11222-012-9348-2
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-012-9348-2