自回归模型参数的经验Bayes估计

Empirical Bayes Estimation of the Parameters About Autoregressive Process

用核估计的方法，构造了自回归模型ＡＲ（ｐ）的参数的一种经验Ｂａｙｅｓ（ＥＢ）估计，证明了这种估计的渐近最优性。模拟计算表明，该估计不仅优于古典估计方法。

The empirical Bayes (EB) estimations of the parameters about autoregresive time series were satdied by Li and Hui (1983), but they did not discuss their asympotically optimality. In this paper we contruct the EB estimators of the parameters for the autoregressive process again. Under a quite general assumption, it is shown that, the estimators are asymptotically optimal EB estimators. Finally we give some simulation results and an empirical comparion with the results both of Li and Hui and of classical methods.