Bayesian Modeling for Regional Sustainable Development

In this paper, a Bayesian network framework has been applied to analyse the probabilistic dynamic of sustainable development at the regional level. The model integrates economic, social, environmental, and governance indicators to capture the uncertainty and independence of factors that conventional deterministic approaches often overlook. Principal nodes, such as economic diversification, innovation capacity, governance quality, energy efficiency, CO2 emissions, social welfare, and a composite sustainable development index form the core structure of the network. Conditional probability tables have been constructed using a hybrid approach that combines statistical data, likelihood-based estimation. Scenario analysis demonstrates that governance quality and innovation capacity exert the strongest influence on sustainability outcomes, while environmental efficiency and emissions reduction provide complementary but less decisive effects. Combined interventions in governance reform, green technology, and innovation yield the highest probability of sustainable development. The findings highlight the risk of a «technological trap» for Russian regions with weak institutions and limited innovation, extending the debate on resource dependency. From a policy perspective, Bayesian networks provide monitoring tools that enable evidencebased decision-making, continuous updating of probability distributions, and better alignment of regional strategies with national goals of sustainability and technological sovereignty.