Estimation via Markov chain Monte Carlo

Proceedings of the American Control Conference - Tập 4 - Trang 2559-2564 vol.4 - 2002
J.C. Spall1
1Applied Physics Laboratory, Johns Hopkins University, Laurel, MD, USA

Tóm tắt

Markov chain Monte Carlo (MCMC) is a powerful means for generating random samples that can be used in computing statistical estimates, numerical integrals, and marginal and joint probabilities. The approach is especially useful in applications where one is forming an estimate based on a multivariate probability distribution or density function that would be hopeless to obtain analytically. In particular, MCMC provides a means for generating samples from joint distributions based on easier sampling from conditional distributions. Over the last 10 to 15 years, the approach has had a large impact on the theory and practice of statistical modeling. On the other hand, MCMC has had relatively little impact (yet) on estimation problems in control. The paper is a survey of popular implementations of MCMC, focusing especially on the two most popular specific implementations of MCMC: Metropolis-Hastings and Gibbs sampling.

Từ khóa

#Monte Carlo methods #Sampling methods #Bayesian methods #Power generation #Books #Physics computing #Probability distribution #Density functional theory #System identification #State estimation

Tài liệu tham khảo

liu, 2001, Monte Carlo Strategies in Scientific Computing 10.1093/biomet/57.1.97 kirkpatrick, 1983, Optimization by simulated annealing, Science, 220, 671, 10.1126/science.220.4598.671 10.1049/ip-f-2.1993.0015 10.1080/00207176908905777 1996, Markov Chain Monte Carlo in Practice 10.1080/01621459.1998.10473766 10.1109/TPAMI.1984.4767596 10.1016/S0167-9473(01)00009-3 parzen, 1962, Stochastic Processes 10.1109/ACC.1997.611067 robert, 1999, Monte Carlo Statistical Methods, 10.1007/978-1-4757-3071-5 10.1016/0005-1098(95)00069-9 10.1109/9.53528 tanner, 1987, The calculation of posterior distributions by data augmentation (with discussion), Journal of the American Statistical Association, 82, 528, 10.1080/01621459.1987.10478458 10.2307/2290282 10.1214/ss/1177010123 10.2307/2289776 10.1109/TAC.1972.1100034 10.2307/2684568 10.1080/01621459.1995.10476635 chen, 2000, Monte Carlo Methods in Bayesian Computation, 10.1007/978-1-4612-1276-8 10.1093/biomet/81.3.541 10.1214/ss/1177009938 10.1109/CDC.2000.912012