Artificial intelligence in the analysis of regional innovation ecosystems of the Russian Federation under import substitution

Abstract

Subject. The article examines the condition of regional innovation ecosystems of the Russian Federation for sustainable development and technological sovereignty of the country under import substitution due to high external challenges. Objectives. The aim is to solve the multifactorial task of analyzing the development of innovative ecosystems in Russia’s regions. The task is characterized by complex formalization, and it is in line with modern concepts of competitive potential, through the proposed productive method, i.e. neural network cluster data analysis. Methods. We employ the cluster analysis based on neural networks that formed an essential component of artificial intelligence. The functionality of artificial neural networks – Kohonen self-organizing maps, involved in this work, has no model prohibitions. The method of neural network cluster analysis enables to visualize clustering results of the multidimensional source data space. Results. The neural network cluster analysis of a set of heterogeneous data helped get the integration of Russian regions across seven clusters. We obtained a significant variety of placement of Russian regions by cluster: the number of regions in clusters varies from one to thirty-one. We established a different level of the state of regional innovation ecosystems, according to the studied indicators in the context of clusters. Conclusions. The findings enabled to assess the state of innovative regional ecosystems in the environment of import substitution created by big challenges from external factors. To continue strengthening sustainable development and technological sovereignty of the country, it is necessary to use different main directions of innovative economic growth of Russian regions, taking into account their specifics in the focus of cluster formations.