Innovative Rating of Regions in the Electronic Industry: Construction and Verification Using Machine Learning

Introduction. The development of the radio-electronic industry is a priority for Russia's technological leadership, necessitating modern tools for assessing the innovative potential of its regions. This study aims to construct and verify an innovative rating of regions for the radio-electronic industry that overcomes the limitations of traditional ratings by applying machine learning to Big Data. Materials and Methods. A training dataset was formed based on Rosstat data from 2010–2022 for 83 regions. Using ensemble machine learning methods (Fine Gaussian SVM, Bagged Trees, Random Forest), a classification model was constructed that assigns innovative ratings (A – leaders, B – average level, C – depressed) to regions based on three target functions, with subsequent aggregation into an integral I-score. A key stage of the research was the model approbation: its verification was carried out on independent data for 2023 that was not part of the training set. Results. The verification confirmed the model's practical applicability: the accuracy of the integral I-score rating prediction on new data was 81.93 %. Based on the approbation results, a current map of innovative ratings was constructed. The leading regions (A) in 2023 were the Moscow Region, Moscow, St. Petersburg, Republic of Tatarstan, Nizhny Novgorod Region, and Sverdlovsk Region. Analysis of discrepancies between prediction and fact revealed growth potential for Novosibirsk Region and potential risks to the leading positions of Republic of Bashkortostan, Perm Territory, and Chelyabinsk Region. Discussion and Conclusion. The approbated methodology enables the construction of accurate and robust assessments of the innovative development of regions in the radio-electronic industry. The verification results demonstrate not only the model's predictive power but also its value for identifying latent trends. The findings are of practical importance for public authorities and large companies in planning regional and sectoral policies.