Search by keyword: machine learning

This study aims to identify groups of Russian regions by their level of sanctions risk prior to the imposition of sanctions and to test the actual exposure to sanctions of these groups after 2022. The application of k-means clustering allowed us to identify three clusters of regions based on their level of sanctions risk using 2021 data: regions with high, medium, and low sanc­tions risk. Data on the number of sanctions imposed on enterprises in the re­gions by the end of 2023 confirmed the validity of the cluster allocation. The average number of sanctions imposed by the end of 2023 on enterprises in regions belonging to the high-sanctions-risk cluster is 129. For the cluster of regions with medium sanctions risk, this figure is 28, and for the low-risk cluster, it is 4. It was found that reliable predictors of a high level of sanctions pressure are such indicators of the regional economy as the region’s enga­gement in trade with overseas countries and the share of the manufacturing industry in the gross regional product. At the same time, economic diver­sification does not always contribute to a lower risk of sanctions pressure.The results of the study have high practical significance for regional econo­mic policy.