目标参数分布情况很复杂，我们想求相关的目标参数（f（x））很难，所以想通过MCMC从目标函数采取样本估计我们想要的结果，大致流程构造一条马尔可夫链去逼近目标函数，从其稳态装下抽取样本。(The distribution of target parameters is very complicated, and it is difficult for us to find the relevant target parameters (f (x)), so we want to use MCMC