Estimation of trends and identification of time series dynamics in short observation sections

被引：1

作者：

Lomov, A. A. ^{[1
]}

机构：

[1] Russian Acad Sci, Inst Math, Siberian Div, Novosibirsk 630090, Russia

来源：

JOURNAL OF COMPUTER AND SYSTEMS SCIENCES INTERNATIONAL | 2009年 / 48卷 / 01期

基金：

俄罗斯基础研究基金会;

关键词：

System Science International; Polynomial Matrix; Toeplitz Matrice; Short Section; Polynomial Matrice;

D O I：

10.1134/S1064230709010018

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Formulas for estimating the trend of a time series by measurements of readings of the summarized "series plus trend" process with additive random disturbances in measurements have been proposed. The series and trend dynamics is described by linear difference equations of a given order. The estimation is based on ensemble measurements of short sections of the summarized process (not longer than transient characteristics). Methods for identifying the parameters of equations and series and trend processes on measurements of the summarized process have been proposed. Applications to the analysis of control systems have been considered.

引用

页码：1 / 13

页数：13