New Method of Order Estimation for ARMA/ARMAX Processes

New Method of Order Estimation for ARMA/ARMAX Processes
复制标题

ARMA/ARMAX 过程阶次估计的新方法

DOI:
10.1137/090768680
复制
发表时间:
2010-03
影响因子:
2.2
通讯作者:
Zhao, Wen-Xiao
Zhao, Wen-Xiao
中科院分区:
数学2区
文献类型:
--
作者:
Chen, Han-Fu;Zhao, Wen-Xiao

文献摘要

参考文献

被引文献

相似文献

设观测值y_k由多元阿尔马过程A(z)y_k=C(z)w_k生成,其系数为$\theta_A$,$\theta_C$,阶数为$(p,r)$,其中w_k $为独立同分布(i.i.d.)具有零均值和未知协方差矩阵R_w>0$的随机向量。本文介绍了一种新的估计阶数(p,r)的方法。与现有的大多数结果相比,新方法不是基于优化某一准则,并且当新数据到达时,与基于准则优化的方法相比,本文给出的阶估计在计算上很容易更新。然后将该方法推广到确定ARMAX过程的阶数。在适当的条件下,证明了当时间趋于无穷大时,估计以概率1收敛到真阶。
Let the observation $\{y_k\}$ be generated by the multivariate ARMA process $A(z)y_k=C(z)w_k$ with unknown coefficients $\theta_A$, $\theta_C$ and orders $(p,r)$, where $\{w_k\}$ is a sequence of independent and identically distributed (i.i.d.) random vectors with zero mean and unknown covariance matrix $R_w>0$. A new method for estimating the orders $(p,r)$ is introduced. In contrast to most of the existing results, the new method is not based on optimizing a certain criterion, and the order estimates given in the paper are rather easy to update computationally in comparison with the criterion-optimization-based methods when new data arrive. The method is then extended to determining the orders of ARMAX processes. Under reasonable conditions the estimates are proved to converge to the true orders with probability one as time tends to infinity.
DOI: 10.1090/cln/007
发表时间: 2001
期刊: --
影响因子: --
作者:
S. Varadhan
通讯作者: S. Varadhan
DOI: 10.1109/tac.2010.2041997
发表时间: 2010-02
影响因子: 6.8
作者:
Han-Fu Chen
通讯作者: Han-Fu Chen
DOI: 10.2307/2988198
发表时间: 1978-12
期刊: The Statistician
影响因子: --
作者:
G. Box;G. Jenkins;G. Reinsel;G. Ljung
通讯作者: G. Box;G. Jenkins;G. Reinsel;G. Ljung
DOI: 10.2307/2983139
发表时间: 1989-05
期刊: --
影响因子: --
作者:
H. Tong;E. Hannan;M. Deistler
通讯作者: H. Tong;E. Hannan;M. Deistler
DOI: 10.1016/b978-008044910-4.00546-0
发表时间: 2009
期刊: --
影响因子: --
作者:
William W. S. Wei
通讯作者: William W. S. Wei