arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.03952stat.MEmath.STstat.COstat.MLstat.TH
收藏

MCMC输出的最优稀疏

Optimal Thinning of MCMC Output

Marina Riabiz, Wilson Chen, Jon Cockayne, Pawel Swietach, Steven A. Niederer, Lester Mackey, Chris. J. Oates

摘要:

使用统计学来评估收敛性和压缩马尔可夫链蒙特卡罗的输出,就所产生的经验近似而言,可能是次优的。通常,许多初始状态归因于“烧入”并被移除,而如果还需要压缩,则链的其余部分被“细化”。在本文中,我们考虑的问题,回顾性地选择一个子集的状态,固定的基数,从样本路径,使其经验分布提供的近似接近最佳。提出了一种新的方法,基于贪婪最小化的内核斯坦差异,这是适合于需要大量压缩的问题。理论结果保证了该方法的一致性,并在常微分方程参数推断的挑战性背景下证明了其有效性。软件在Stein Thinning包中可用Python,R和MATLAB。

英文摘要:

The use of heuristics to assess the convergence and compress the output of Markov chain Monte Carlo can be sub-optimal in terms of the empirical approximations that are produced. Typically a number of the initial states are attributed to "burn in" and removed, whilst the remainder of the chain is "thinned" if compression is also required. In this paper we consider the problem of retrospectively selecting a subset of states, of fixed cardinality, from the sample path such that the approximation provided by their empirical distribution is close to optimal. A novel method is proposed, based on greedy minimisation of a kernel Stein discrepancy, that is suitable for problems where heavy compression is required. Theoretical results guarantee consistency of the method and its effectiveness is demonstrated in the challenging context of parameter inference for ordinary differential equations. Software is available in the Stein Thinning package in Python, R and MATLAB.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑