吉布斯(Gibbs)抽样方法是 Markov Chain Monte Carlo(MCMC)方法的一种,也是应用最为广泛的一种(The simplest Gibbs sampling is a special case of Metropolis-Hastings algorithm, while the extension of Gibbs sampling can be regarded as a universal sampling system. This system takes a sample of each (or each) variable by rotation and combines a Metropolis-Hastings algorithm (or more complex algorithms, such as slice sampling, adaptive rejection sampling, and adaptive rejection Metropolis algorithm) to take a step or multistep sampling of a large number of variables.)