Typology of Russian regions by the functions of key settlements: a system of geospatial artificial intelligence

The purpose of the study is to develop a geospatial artificial intelligence system and use it to identify types and other taxonomic units as groups of Russian regions with a specific combination of functions for key settlements. The list of such settlements and the criteria for their selection (functions) have been approved by the Government of the Russian Federation. Each region has a specific combination of key settlements that indicates the region's place and role in Russia's spatial development. To identify this role, which is not documented in government strategies or scientific publications, a computational system combining machine learning, random forest methods, and inference rules was developed. Within the regions, the distribution of key settlements between the center, semi-periphery, and periphery was carried out, followed by the typology of the regions. The result was the division of 89 regions into 3 types, 18 subtypes, and 52 kinds. It has been established that Russia has a special territorial structure of regional combinations of key settlements’ functions. This structure can be represented as a core (a compact set of regions with a complete center-peripheral structure) and a discontinuous periphery with an incomplete structure. The practical significance of the results may be related to the adjustment to the state’s spatial development strategy and the refinement of the key settlements’ list in Russia.