zbMATH — the first resource for mathematics

Automatically tuned general-purpose MCMC via new adaptive diagnostics. (English) Zbl 1417.65017
Summary: Adaptive Markov Chain Monte Carlo (MCMC) algorithms attempt to ‘learn’ from the results of past iterations so the Markov chain can converge quicker. Unfortunately, adaptive MCMC algorithms are no longer Markovian, so their convergence is difficult to guarantee. In this paper, we develop new diagnostics to determine whether the adaption is still improving the convergence. We present an algorithm which automatically stops adapting once it determines further adaption will not increase the convergence speed. Our algorithm allows the computer to tune a ‘good’ Markov chain through multiple phases of adaption, and then run conventional non-adaptive MCMC. In this way, the efficiency gains of adaptive MCMC can be obtained while still ensuring convergence to the target distribution.
65C05 Monte Carlo methods
65C40 Numerical analysis or methods applied to Markov chains
60J22 Computational methods in Markov chains
60J05 Discrete-time Markov processes on general state spaces
62F15 Bayesian inference
Full Text: DOI
[1] Andrieu, C; Atchadé, YF, On the efficiency of adaptive MCMC algorithms, Electron Commun Probab, 12, 336-349, (2007) · Zbl 1129.65006
[2] Andrieu, C; Moulines, E, On the ergodicity properties of some adaptive Markov chain Monte Carlo algorithms, Ann Appl Probab, 16, 1462-1505, (2006) · Zbl 1114.65001
[3] Atchadé, YF; Rosenthal, JS, On adaptive Markov chain Monte Carlo algorithms, Bernoulli, 11, 815-828, (2005) · Zbl 1085.62097
[4] Brooks S, Gelman A, Jones GL, Meng X (eds) (2011) Handbook of Markov Chain Monte Carlo. Taylor & Francis, Boca Raton · Zbl 1218.65001
[5] Brooks, SP; Gelman, A, General methods for monitoring convergence of iterative simulations, J Comput Graph Stat, 7, 434-455, (1998)
[6] Fort, G; Moulines, E; Priouret, P, Convergence of adaptive and interacting Markov chain Monte Carlo algorithms, Ann Stat, 39, 3262-3289, (2011) · Zbl 1246.65003
[7] Gaver, DP; O’Muircheartaigh, IG, Robust empirical Bayes analyses of event rates, Technometrics, 29, 1-15, (1987) · Zbl 0611.62124
[8] Gelfand, AE; Smith, AFM, Sampling-based approaches to calculating marginal densities, J Am Stat Assoc, 85, 398-409, (1990) · Zbl 0702.62020
[9] Gelman, A; Roberts, GO; Gilks, WR; Bernardo, JM (ed.); etal., Efficient metropolis jumping rules, No. 5, 599-607, (1996), Oxford
[10] Gelman, A; Rubin, DB, Inference from iterative simulation using multiple sequences, Stat Scie, 7, 457-472, (1992) · Zbl 1386.65060
[11] George, EI; Makov, UE; Smith, AFM, Conjugate likelihood distributions, Scand J Stat, 20, 147-156, (1993) · Zbl 0776.62027
[12] Geyer CJ, Johnson LT (2014) MCMC: Markov Chain Monte Carlo. R package version 0.9-3. http://CRAN.R-project.org/package=mcmc
[13] Giordani, P; Kohn, R, Adaptive independent metropolis-Hastings by fast estimation of mixtures of normals, J Comput Graph Stat, 19, 243-259, (2010)
[14] Haario, H; Laine, M; Mira, A; Saksman, E, DRAM: efficient adaptive MCMC, Stat Comput, 16, 339-354, (2006)
[15] Haario, H; Saksman, E; Tamminen, J, An adaptive metropolis algorithm, Bernoulli, 7, 223-242, (2001) · Zbl 0989.65004
[16] Hastings, WK, Monte Carlo sampling methods using Markov chains and their applications, Biometrika, 57, 97-109, (1970) · Zbl 0219.65008
[17] Lunn D, Spiegelhalter D, Thomas A, Best N (2009) The BUGS project: evolution, critique, and future directions. Stat Med 28(25):3049-3067. http://www.openbugs.net · Zbl 0776.62027
[18] Metropolis, N; Rosenbluth, AW; Rosenbluth, MN; Teller, AH; Teller, E, Equation of state calculations by fast computing machines, J Chem Phys, 21, 1087-1092, (1953)
[19] Neal, PJ; Roberts, GO; Yuen, WK, Optimal scaling of random walk metropolis algorithms with discontinuous target densities, Ann Appl Probab, 22, 1880-1927, (2012) · Zbl 1259.60082
[20] R Core Team (2014) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org · Zbl 0702.62020
[21] Roberts, GO; Gelman, A; Gilks, WR, Weak convergence and optimal scaling of random walk metropolis algorithms, Ann Appl Probab, 7, 110-120, (1997) · Zbl 0876.60015
[22] Roberts, GO; Rosenthal, JS, Optimal scaling for various metropolis-Hastings algorithms, Stat Sci, 16, 351-367, (2001) · Zbl 1127.65305
[23] Roberts, GO; Rosenthal, JS, General state space Markov chains and MCMC algorithms, Probab Surv, 1, 20-71, (2004) · Zbl 1189.60131
[24] Roberts, GO; Rosenthal, JS, Coupling and ergodicity of adaptive Markov chain Monte Carlo algorithms, J Appl Probab, 44, 458-475, (2007) · Zbl 1137.62015
[25] Roberts, GO; Rosenthal, JS, Examples of adaptive MCMC, J Comput Graph Stat, 18, 349-367, (2009)
[26] Rosenthal JS (2004) Adaptive MCMC Java applet. http://probability.ca/jeff/java/adapt.html · Zbl 1386.65060
[27] Rosenthal, JS, AMCMC: an R interface for adaptive MCMC, Comput Stat Data Anal, 51, 5467-5470, (2007) · Zbl 1203.62001
[28] Rosenthal JS (2007b) The AMCMC package. http://probability.ca/amcmc · Zbl 1114.65001
[29] Rosenthal, JS; Brooks, S (ed.); Gelman, A (ed.); Jones, GL (ed.); Meng, X (ed.), Optimal proposal distributions and adaptive MCMC, 93-112, (2011), Boca Raton · Zbl 1229.65021
[30] Scheidegger A (2012) adaptMCMC: Implementation of a generic adaptive Monte Carlo Markov Chain sampler. R package version 1.1. http://CRAN.R-project.org/package=adaptMCMC
[31] Soetaert, K; Petzoldt, T, Inverse modelling, sensitivity and Monte Carlo analysis in R using package FME, J Stat Softw, 33, 1-28, (2010)
[32] Soetaert K, Petzoldt T (2014) FME: a flexible modelling environment for inverse modelling, sensitivity, identifiability, Monte Carlo analysis. R package version 1.3.1. http://CRAN.R-project.org/package=FME · Zbl 0876.60015
[33] Turro, E; Bochkina, N; Hein, AMK; Richardson, S, BGX: a bioconductor package for the Bayesian integrated analysis of affymetrix genechips, BMC Bioinform, 8, 439-448, (2007)
[34] Vihola, M, Grapham: graphical models with adaptive random walk metropolis algorithms, Comput Stat Data Anal, 54, 49-54, (2010) · Zbl 1284.62037
[35] Vihola M (2010b) The Grapham package. http://www.stats.ox.ac.uk/ mvihola/grapham/ · Zbl 1284.62037
[36] Vihola, M, Robust adaptive metropolis algorithm with coerced acceptance rate, Stat Comput, 22, 997-1008, (2012) · Zbl 1252.65024
[37] Yang J (2014) atmcmc: Automatically tuned Markov Chain Monte Carlo. R package version 1.0.http://CRAN.R-project.org/package=atmcmc
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.