Componentwise adaptation for high dimensional MCMC

被引：158

作者：

Haario, H

Saksman, E

Tamminen, J

机构：

[1] Univ Helsinki, Dept Math & Stat, FIN-00014 Helsinki, Finland

[2] Univ Jyvaskyla, Dept Math & Stat, FIN-40014 Jyvaskyla, Finland

[3] Finnish Meteorol Inst, Geophys Res Div, FIN-00101 Helsinki, Finland

来源：

COMPUTATIONAL STATISTICS | 2005年 / 20卷 / 02期

关键词：

MCMC; adaptive MCMC; Metropolis-Hastings algorithm;

D O I：

10.1007/BF02789703

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

We introduce a new adaptive MCMC algorithm, based on the traditional single component Metropolis-Hastings algorithm and on our earlier adaptive Metropolis algorithm (AM). In the new algorithm the adaption is performed component by component. The chain is no more Markovian, but it remains ergodic. The algorithm is demonstrated to work well in varying test cases up to 1000 dimensions.

引用

页码：265 / 273

页数：9

共 11 条

[1]

Andrieu C., 2001, WORKING PAPERS, P2001

[2]

[Anonymous], ADAPTIVE MARKOV CHAI

[3]

GELMAN A, 1996, BAYESIAN STAT, V5, P59

[4] Adaptive Markov chain Monte Carlo through regeneration [J].

Gilks, WR ;

Roberts, GO ;

Suhu, SK .

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1998, 93 (443) :1045-1054

[5]

GILKS WR, 1995, MARKOV CHAIN MONTE C, P75

[6] An adaptive Metropolis algorithm [J].

Haario, H ;

Saksman, E ;

Tamminen, J .

BERNOULLI, 2001, 7 (02) :223-242

[7] Adaptive proposal distribution for random walk Metropolis algorithm [J].

Haario, H ;

Saksman, E ;

Tamminen, J .

COMPUTATIONAL STATISTICS, 1999, 14 (03) :375-395

[8]

HAARIO H, 2003, 342 U HELS DEP MATH

[9]

HASTINGS WK, 1970, BIOMETRIKA, V57, P97, DOI 10.1093/biomet/57.1.97

[10] EQUATION OF STATE CALCULATIONS BY FAST COMPUTING MACHINES [J].

METROPOLIS, N ;

ROSENBLUTH, AW ;

ROSENBLUTH, MN ;

TELLER, AH ;

TELLER, E .

JOURNAL OF CHEMICAL PHYSICS, 1953, 21 (06) :1087-1092

← 1 2 →