Adaptive Metropolis algorithm using variational Bayesian adaptive Kalman filter

被引：37

作者：

Mbalawata, Isambi S. ^{[1
]}

Sarkka, Simo ^{[2
]}

Vihola, Matti ^{[3
]}

Haario, Heikki ^{[1
]}

机构：

[1] Lappeenranta Univ Technol, Dept Math & Phys, FI-53851 Lappeenranta, Finland

[2] Aalto Univ, Dept Biomed Engn & Computat Sci, FI-00076 Aalto, Finland

[3] Univ Oxford, Dept Stat, Oxford OX1 3TG, England

来源：

COMPUTATIONAL STATISTICS & DATA ANALYSIS | 2015年 / 83卷

基金：

芬兰科学院;

关键词：

Markov chain Monte Carlo; Adaptive Metropolis algorithm; Adaptive Kalman filter; Variational Bayes; CHAIN MONTE-CARLO; ERGODICITY; SIMULATION;

D O I：

10.1016/j.csda.2014.10.006

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

Markov chain Monte Carlo (MCMC) methods are powerful computational tools for analysis of complex statistical problems. However, their computational efficiency is highly dependent on the chosen proposal distribution, which is generally difficult to find. One way to solve this problem is to use adaptive MCMC algorithms which automatically tune the statistics of a proposal distribution during the MCMC run. A new adaptive MCMC algorithm, called the variational Bayesian adaptive Metropolis (VBAM) algorithm, is developed. The VBAM algorithm updates the proposal covariance matrix using the variational Bayesian adaptive Kalman filter (VB-AKF). A strong law of large numbers for the VBAM algorithm is proven. The empirical convergence results for three simulated examples and for two real data examples are also provided. (C) 2014 Elsevier B.V. All rights reserved.

引用

页码：101 / 115

页数：15

共 39 条

[1] On the ergodicity properties of some adaptive MCMC algorithms
Andrieu, Christophe
Moulines, Eric
[J]. ANNALS OF APPLIED PROBABILITY, 2006, 16 (03) : 1462 - 1505
[2] A tutorial on adaptive MCMC
Andrieu, Christophe
Thoms, Johannes
[J]. STATISTICS AND COMPUTING, 2008, 18 (04) : 343 - 373
[3] [Anonymous], 2004, Estimation with applications to tracking and navigation: theory algorithms and software
[4] [Anonymous], 2007, Stochastic processes and filtering theory
[5] On adaptive Markov chain Monte Carlo algorithms
Atchadé, YF
Rosenthal, JS
[J]. BERNOULLI, 2005, 11 (05) : 815 - 828
[6] Limit theorems for some adaptive MCMC algorithms with subgeometric kernels
Atchade, Yves
Fort, Gersende
[J]. BERNOULLI, 2010, 16 (01) : 116 - 154
[7] Bai Y., 2009, THESIS U TORONTO
[8] Bai Y, 2011, ADV APPL STAT, V21, P1
[9] Berthouex P., 2002, STAT ENV ENG
[10] Approximating Probability Densities by Iterated Laplace Approximations
Bornkamp, Bjoern
[J]. JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS, 2011, 20 (03) : 656 - 669

← 1 2 3 4 →