MCMC Algorithms for Posteriors on Matrix Spaces

MCMC Algorithms for Posteriors on Matrix Spaces
复制标题

矩阵空间上后验的 MCMC 算法

DOI:
10.1080/10618600.2022.2058953
复制
发表时间:
2022
影响因子:
2.4
通讯作者:
Kamatani Kengo
Kamatani Kengo
中科院分区:
数学2区
文献类型:
--
作者:
Beskos Alexandros;Kamatani Kengo

文献摘要

参考文献

相似文献

研究了目标分布在矩阵空间上的马尔可夫链蒙特卡罗(MCMC)算法。这样一个重要的抽样问题还有待分析探讨。我们进行了一个重要的步骤,通过开发适当的理论框架,允许识别遍历性的典型MCMC算法,在这样的背景下相关的属性,以覆盖这一差距。除了标准的随机游走大都会(RWM)和预处理的Crank-Nicolson(pCN),本文的贡献,在一个新的算法,称为“混合”pCN(MpCN)的发展。对于一类重要的重尾矩阵分布,证明了RWM和pCN不是几何遍历的。相比之下,MpCN在具有不同尾部行为的目标之间是稳健的,并且在重尾分布类内具有非常好的经验性能。MpCN的几何遍历性在这项工作中没有得到充分证明,因为由于状态空间的复杂性,一些剩余的漂移条件很难获得。然而,我们确实在证明方面取得了很大的进展,并详细展示了为未来工作留下的最后步骤。我们通过数值应用说明了各种算法的计算性能,包括在金融统计中产生的具有挑战性的模型的真实的数据上的校准。本文的补充材料可在网上查阅。
We study Markov chain Monte Carlo (MCMC) algorithms for target distributions defined on matrix spaces. Such an important sampling problem has yet to be analytically explored. We carry out a major step in covering this gap by developing the proper theoretical framework that allows for the identification of ergodicity properties of typical MCMC algorithms, relevant in such a context. Beyond the standard Random-Walk Metropolis (RWM) and preconditioned Crank–Nicolson (pCN), a contribution of this article in the development of a novel algorithm, termed the “Mixed” pCN (MpCN). RWM and pCN are shownnotto be geometrically ergodic for an important class of matrix distributions with heavy tails. In contrast, MpCN is robust across targets with different tail behavior and has very good empirical performance within the class of heavy-tailed distributions. Geometric ergodicity for MpCN is not fully proven in this work, as some remaining drift conditions are quite challenging to obtain owing to the complexity of the state space. We do, however, make a lot of progress toward a proof, and show in detail the last steps left for future work. We illustrate the computational performance of the various algorithms through numerical applications, including calibration on real data of a challenging model arising in financial statistics. Supplementary materials for this article are available online.
DOI: 10.1214/ecp.v18-2507
发表时间: 2012-12
影响因子: 0.5
作者:
Daniel Rudolf;Mario Ullrich
通讯作者: Daniel Rudolf;Mario Ullrich
DOI: 10.4171/rlm/694
发表时间: 2014
影响因子: 0.5
作者:
A. Kulik;M. Scheutzow
通讯作者: M. Scheutzow
DOI: 10.1239/jap/1085496589
发表时间: 2001-02
影响因子: 1
作者:
G. Roberts;R. Tweedie
通讯作者: G. Roberts;R. Tweedie
DOI: --
发表时间: 1967
期刊:
影响因子: --
作者:
J. Dickey
通讯作者: J. Dickey
DOI: 10.1198/016214508000000724
发表时间: 2008-12-01
影响因子: 3.7
作者:
O'Malley, A. James;Zaslavsky, Alan M.
通讯作者: Zaslavsky, Alan M.