Open AccessIdentification of an autoregressive process using Matlab
Author(s) -
А. Sherstneva,
SibSUTIS,
O. Sherstneva
Publication year - 2021
Publication title -
t-comm
Language(s) - English
Resource type - Journals
eISSN - 2072-8743
pISSN - 2072-8735
DOI - 10.36724/2072-8735-2021-15-7-34-38
Subject(s) - computer science , identification (biology) , matlab , reliability (semiconductor) , autoregressive model , process (computing) , task (project management) , data mining , regression analysis , estimation theory , system identification , autoregressive–moving average model , time series , machine learning , algorithm , statistics , mathematics , engineering , power (physics) , botany , physics , systems engineering , quantum mechanics , biology , measure (data warehouse) , operating system
The article aims to apply the methods of regression analysis in estimation task the reliability indicators of an infocommunication system and data trend forecasting. The task of assessment the reliability indicators of infocommunication systems is based, traditionally, on statistical data, the collection and processing is carried out by monitoring system. To obtain the calculated indicators as close as possible to the real practical results it is necessary to process a large number of measurements. In this sense, the theory of filtration is widely used in various estimation problems. It allows to support the possibility of effective solution of technical issues and implementation through mathematical modeling programs. The article is aimed to consider the issues of data trend forecasting for parameters calculation of infocommunication systems. One of the most effective solutions is the use of regression analysis methods. The research gives development and analysis of mathematical and program models. Based on theoretical calculations, methods for solving the problem are determined and experimental research are carried out. The article solves the problem of identifying an autoregressive process through the Yule-Walker equations. In addition to theoretical calculations, a program has been developed that allows handle the process of calculations in the Matlab environment. As a result, it is proposed to compare the results of identification time series of variables through the Yule-Walker system with the classical parameter estimation by the least squares method. The results are shown graphically.