数据缺失条件下AR(p)模型参数不确定性研究

The research of model parameters uncertainty under absence data conditions

在数据缺失条件下,探讨贝叶斯推断法在AR(p)模型参数不确定性研究中的应用。根据Bayesian理论与Markov chain Monte Carlo(简称MCMC)法,在WinBUGS软件中搜索一个或更多数据缺失时AR(p)模型参数的后验状态空间。计算结果表明,贝叶斯推断法既能充分利用现有资料信息,又能考虑参数的不确定性,是一种AR(p)模型参数估计的有效方法。

under the condition of the absence of data,the Baysian inference method was discussed on the appliction of uncertainty research of the AR（p） model parameter.According to Bayesian theory and Markov chain Monte Carlo（abbreviated MCMC） method,it searcheed one or more absence data with the WinBUGS software when the AR（p） model parameters posterior state space.The results showed that Bayesian inference method can not only fully utilize existing data and information,but also consider the uncertainty of parameters.It is an effective method of AR（p） model parameter estimation.