一类双重时序模型AR(1)-MA(0)的参数矩估计及其渐近性质  

Moment Estimation of Parameters and its Asymptotic Properties about a Class of Doubly Time Scries Models AR(1)-MA(0)

作　　者：苗夺谦[1] 常学将 

机构地区：[1]山西大学

出　　处：《工程数学学报》1991年第3期73-82,共10页Chinese Journal of Engineering Mathematics

摘　　要：近年来,人们将时间序列的研究重点,逐渐转向非线性时序模型,双重时序模型就是一类很重要的非线性模型,从目前文献来看,对非线性模型的研究主要是讨论平稳解存在的条件及模型预报问题;对于模型参数的估计及其渐近性质很少研究。 本文利用矩方法,给出了双重时序模型AR(1)-MA(O)的参数矩估计。在第二重模型噪声方差已知的条件下,通过对协方差函数渐近性质的研究,证明了该估计的相容性和渐近正态性。Recently, on the reserch of time series people put more attention to non-linear models. Doubly Stochastic time Scries model is one class of very important non-linear models. In view of present papers, stationarity and forecasting of non-linear models are mainly studied, but estimation of parameters and its limit behavier have rarely been discussed.In this paper, we give a moment estimation of parameters of AR(1)-MA(0) model, using moment method. Given the variance of noise in second layer model, we prove the consistence and asymptotic normality.

关 键 词：双重时序模型 参数矩估计 矩方法 

分 类 号：O211.61[理学—概率论与数理统计]