Clustering of multidimensional temporal data as part of information support for management decisions

Objectives . The aim of information support for management decision-making is to find the most optimal option. Cluster analysis of multivariate data characterizing socioeconomic systems is widely used. In this work, the author aims to increase the efficiency of decisions made to manage these systems based on the clustering of temporal multidimensional data. Methods . The methods of cluster analysis were used, as well as the provisions of the theory of systems and mathematical statistics. Results. A methodology for analyzing the functioning of socioeconomic systems was developed. The analysis is implemented in three stages. Firstly, clustering over the values of feature variances was applied. Secondly, the distance of clustering objects from the center of their cluster and their dispersion was calculated at the points of time coordinates. Thirdly, the change in belonging to a certain cluster of objects that came into view earlier was monitored. Unstable systems were then identified. Conclusions . Two cases were considered to justify the effectiveness of the methodology developed herein. First, using the example of the tax administration, the detection of deliberate distortion of information was considered. Secondly, identifying the abnormal functioning of the regions of the Russian Federation using the example of decision-making in the framework of socioeconomic development management was considered. The analysis demonstrated good results and we can thus recommend the proposed methodology for practical use in information systems for supporting management decisions.