Markov Chain Monte Carlo Algorithms
Markov Chain Monte Carlo Algorithms
复制标题
DOI：
10.1007/978-4-431-55060-0_26
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
O. Maruyama
中科院分区：
文献类型：
--
作者：
O. Maruyama
马尔可夫链蒙特卡罗（MCMC）方法是一种从特定概率分布生成样本的通用算法框架。当从分布中直接抽样是未知的时候，它们是有用的。本文介绍了MCMC的理论，介绍了两种典型的MCMC算法（Metropolis-Hastings和Gibbs采样）和三种回火方法（模拟回火、并行回火和模拟退火），并讨论了MCMC方法在系统生物学预测问题中的应用。
Markov chain Monte Carlo (MCMC) methods are a general framework of algorithms for generating samples from a specified probability distribution. They are useful when direct sampling from the distribution is unknown. This article describes theory of MCMC, presents two typical MCMC algorithms (Metropolis-Hastings and Gibbs sampling) and three tempering methods (simulated tempering, parallel tempering, and simulated annealing), and discusses the application of MCMC methods to a prediction problem in systems biology.