Кластерный анализ пространственно-временных паттернов в исследовании структуры деловой активности региона

Abstract

The paper is devoted to the urgent task of studying the business activity of business in a particular territory. It examines changes in the effectiveness of doing business in a certain region depending on the time, location, and economic conditions of conducting business activities of small and medium-sized enterprises. A formulated methodological approach allows analyzing and identifying patterns in the distribution of performance across areas of activity of entities, taking into account spatial and temporal aspects. Cluster analysis methods are used to identify general patterns and factors affecting economic development, as well as to develop recommendations for optimizing business processes and improving the conditions for entrepreneurial activity at the regional level. The effectiveness of solving the set task in the analysis of complex high-dimensional data using nonlinear methods of dimensionality reduction is shown. The results of the research showed that business activity in the region depends not only on economic factors observed in dynamics, but also on the spatial location of economic entities, the clustering of which is performed using specialized metrics, which allows adapting the universal UMAP feature space dimensionality reduction algorithm to specific tasks and data types. The findings may be useful for entrepreneurs, authorities, and business communities seeking to optimize business support measures and improve conditions for doing business.