The purpose of this study is to develop and test an economicmathematical toolkit for segmenting the regions of the Russian Federation based on institutional and innovation indicators, with the subsequent identification of key determinants of cluster formation. The paper presents a methodological approach to the classification of Russian regions using institutional and innovationrelated metrics. Regional segmentation was carried out using the kmeans algorithm and hierarchical agglomerative clustering. To assess the influence of key factors, multiple linear regression was applied. The density of small enterprises and the volume of capital investment showed a statistically significant negative association with cluster affiliation at the 0,01 significance level. Internal expenditures on research and development demonstrated a positive contribution, whereas the output of innovative products revealed a negative effect at the 0,05 significance level. The hierarchical model additionally confirmed the statistical relevance of the number of R&D personnel and patent activity. The coefficient of determination for the models was R² = 0,676 and R² = 0,785, respectively. To verify the robustness of the resulting clusters, classification models were employed, including Random Forest, Support Vector Machines, kNearest Neighbors, and Gradient Boosting. Random Forest achieved the highest classification accuracy (96%) and a mean F1 score exceeding 0,95. The results obtained provide an empirical foundation for the development of targeted strategies to support innovationdriven regional development.
Зайцев et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: