Artificial intelligence methods in the study of the economic potential of Russian regions in conditions of grand challenges

Abstract

Subject. The article investigates methodological approaches to the analysis of economic potential of regions, considering the achievement of the national goal of sustainable development of the Russian Federation in conditions of grand challenges. Objectives. The aim is to study the dynamics of economic activity in Russian regions, using artificial intelligence methods, to analyze the innovative development of the Russian economy in the face of grand challenges. Methods. The study rests on the analysis of development indicators of the regional economy of Russia. We propose a cluster analysis of the regional economy development, free from model constraints, based on neural network modeling, which enables to assess the dynamics of development and ranking of Russian regions, according to the totality of considered indicators. We apply Kohonen self-organizing maps as a promising means of clustering and visual embodiment of multidimensional statistical data. Results. The neural network modeling enabled to segregate 85 regions of the Russian Federation into four compact groups. We estimated the significance of each indicator in the formation of clusters, revealed a strong difference in the number of regions in the clusters. In the period under review, some regions were part of the same corresponding cluster. The paper presents the dynamics of average values of the studied indicators in clusters for 2018–2020. Conclusions. We demonstrate a disproportion of economic development of Russian regions. It requires an individual approach to regional economy’s strategy development, corresponding KPIs, and measures to stimulate economic activity in the field of innovation, investment, and introduction of research results in the regions of the Russian Federation.