UDC 004.724
METHOD TO IDENTIFY STRUCTURAL TRANSFORMATIONS OF TIME SERIES USING FUZZY CLUSTERING PRINCIPLES
M. A. Stepanov, graduate student RSREU, Ryazan, Russia;
orcid.org/0000-0003-1165-415X, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The problem of identifying structural transformations in time series groups has been considered. The aim of the work is to develop the method for identifying large, medium and small transformations in the groups of time series, based on the principles of fuzzy clustering. The algorithm for identifying large, medium, and small transformations in time series groups is proposed. It implements the making-decision based on a comparative analysis of the clustering results using a fuzzy c-means algorithm for time series groups whose length differs by one. An approach to the presentation of time series groups used in the clustering procedure, based on their trends, as well as an approach to the calculation of relevance coefficients of the elements of group time series are proposed. The examples of identifying structural transformations in the
problem of analyzing the group of model time series, as well as the group of time series characterizing the regional socio-economic sphere, confirming the effectiveness of the proposed method have been given.
Key words: time series, fuzzy clustering, optimal number of clusters, identification of the structural transformations of time series.