A recursive approach to time-series analysis for multi-variable systems
A recursive approach to time-series analysis for multi-variable systems
复制标题
多变量系统时间序列分析的递归方法
DOI：
10.1080/00207177708922245
复制
发表时间：
1977
影响因子：
2.1
通讯作者：
P. Whitehead
中科院分区：
文献类型：
--
作者：
P. Young;P. Whitehead
本文将单输入、单输出系统的时间序列分析的递归工具变量近似最大似然（IV-AML）方法扩展到多变量（多输入、多输出系统）的表征，使用统计特征选择技术以近似的方式规避表征多变量随机扰动的一些问题。还概述了一种基于多变量 IV-AML 程序的动态系统结构识别和参数估计方法，并将其应用于基于长期收集的日常现场数据对非潮汐河流系统中生化需氧量 (BOD) 和溶解氧 (DO) 之间的动态关系进行建模的问题。
In this paper the recursive instrumental variable-approximate maximum likelihood (IV-AML) method of time-series analysis for single-input, single-output systems is extended to the characterization of multivariable (multi-input, multi-output systems) using techniques of statistical feature selection to circumvent, in an approximate fashion, some of the problems of characterizing multi-variable stochastic disturbances. A method of dynamic system structure identification and parameter estimation based on this multi-variable IV-AML procedure is also outlined and applied to the problem of modelling the dynamic relationship between biochemical oxygen demand (BOD) and dissolved oxygen (DO) in a non-tidal river system on the basis of daily field data collected over an extended period of time.