Componentwise adaptation for high dimensional MCMC
| 详细信息 | 推荐本文 | |
- 作者：Heikki Haario (1)
Eero Saksman (2)
Johanna Tamminen (3) - 关键词：MCMC ; adaptive MCMC ; Metropolis ; Hastings algorithm
- 刊名：Computational Statistics
- 出版时间：June 2005
- 出版年：2005
- 期刊代码：pc577_0943-4062
- 类别：mat
- 卷：20
- 期：2
- 页码：265-273
- 数据来源：sp
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.