基于贝叶斯推断的缓冲自回归模型参数估计

Parameter Estimation of Buffered Autoregressive Model Based on Bayesian Inference

摘要: 本文基于马尔科夫链蒙特卡罗(MCMC)算法的贝叶斯推断下，研究缓冲自回归模型的参数估计问题。通过缓冲自回归模型参数的联合后验分布得到各参数的条件后验分布，再利用Gibbs抽样、随机游走Metropolis-Hastings算法抽取样本，并以正态分布为建议分布，对缓冲自回归模型进行参数估计。随机模拟结果显示用该方法估计各参数的效果较好。

Abstract: Based on Bayesian inference of Markov chain Monte Carlo (MCMC) algorithm, this paper studies the parameter estimation of buffered autoregressive models. The conditional posterior distribution of each parameter is obtained through the joint posterior distribution of the parameters of the buffered autoregressive model. Then, the Gibbs sampling and random walk Metropolis Hastings algorithm are used to extract samples, and the normal distribution is used as the suggested distribution to estimate the parameters of the buffered autoregressive model. The results of stochastic simulation show that the method is effective in estimating the parameters.