Marginal reversible jump Markov chain Monte Carlo with application to motor unit number estimation
Marginal reversible jump Markov chain Monte Carlo with application to motor unit number estimation
复制标题
DOI：
10.1016/j.csda.2013.11.003
复制
发表时间：
2014-04-01
影响因子：
1.8
通讯作者：
McCombe, Pamela A.
中科院分区：
文献类型：
--
作者：
Drovandi, Christopher C.;Pettitt, Anthony N.;McCombe, Pamela A.
运动单元数估计（Motor unit number estimation， MUNE）是一种旨在为导致运动神经元疾病等运动单元丧失的疾病的进展提供定量指标的方法。然而，开发一种可靠的、可重复的、快速的实时MUNE方法一直是一个难题。在此之前，一种可逆跳跃马尔可夫链蒙特卡罗（RJMCMC）算法已经实现，该算法使用贝叶斯分层模型产生运动单元数量的后验分布，该模型考虑了运动单元激活的生物信息。然而，这种方法对于某些数据集可能不可靠，因为它可能受到跨维混合不良的影响。重点是通过边缘化潜在变量来创建可能性来改进推理。更具体地说，重点是这种边缘化如何改善RJMCMC混合，以及如何研究利用可能性的替代方法（例如DIC）。对于该模型，边缘化是对潜在变量的，对于大量的运动单元，是对一组潜在二进制变量的所有组合的难以处理的求和，这些变量的联合样本空间随着运动单元的数量呈指数增长。提供了这个数量的一个易于处理和准确的近似值，并研究了其他基于蒙特卡罗估计的近似值，这些近似值可以纳入RJMCMC。(C) 2013 Elsevier B.V.版权所有
Motor unit number estimation (MUNE) is a method which aims to provide a quantitative indicator of progression of diseases that lead to a loss of motor units, such as motor neurone disease. However the development of a reliable, repeatable and fast real-time MUNE method has proved elusive hitherto. Previously, a reversible jump Markov chain Monte Carlo (RJMCMC) algorithm has been implemented to produce a posterior distribution for the number of motor units using a Bayesian hierarchical model that takes into account biological information about motor unit activation. However this approach can be unreliable for some datasets since it can suffer from poor cross-dimensional mixing. The focus is on improved inference by marginalising over latent variables to create the likelihood. More specifically, the emphasis is on how this marginalisation can improve the RJMCMC mixing and that alternative approaches that utilise the likelihood (e.g. DIC) can be investigated. For this model the marginalisation is over latent variables which, for a larger number of motor units, is an intractable summation over all combinations of a set of latent binary variables whose joint sample space increases exponentially with the number of motor units. A tractable and accurate approximation for this quantity is provided and also other approximations based on Monte Carlo estimates that can be incorporated into RJMCMC are investigated. (C) 2013 Elsevier B.V. All rights reserved.