Adaptive Gibbs samplers and related MCMC methods
Adaptive Gibbs samplers and related MCMC methods
复制标题
DOI：
10.1214/11-aap806
复制
发表时间：
2011-01
影响因子：
1.8
通讯作者：
K. Latuszy'nski;G. Roberts;J. Rosenthal
中科院分区：
文献类型：
--
作者：
K. Latuszy'nski;G. Roberts;J. Rosenthal
我们考虑各种版本的自适应吉布斯采样器和吉布斯内的大都会采样器，它们在运行期间通过学习来更新它们在 y 上的选择概率（也许还有它们的建议分布），以尝试优化算法。我们提出了一个警示性的例子，说明即使是看似简单的自适应吉布斯采样器也可能无法收敛。然后，我们提出了各种积极的结果，保证了自适应吉布斯采样器在某些条件下的收敛。 AMS 2000 学科分类：初级 60J05、65C05；次级 62F15。
We consider various versions of adaptive Gibbs and Metropolis- within-Gibbs samplers, which update their selection probabilities (and per- haps also their proposal distributions) on the y during a run, by learning as they go in an attempt to optimise the algorithm. We present a cautionary example of how even a simple-seeming adaptive Gibbs sampler may fail to converge. We then present various positive results guaranteeing convergence of adaptive Gibbs samplers under certain conditions. AMS 2000 subject classications: Primary 60J05, 65C05; secondary 62F15.