Expansion of Methodological Approaches to Assessing the Stability of Commercial Banks by Artifi cial Intelligence Systems

The article considers the issues of sustainable development of commercial banks taking into account compliance with the ESG principles of the green economy. The relevance is that in modern conditions, which are characterized by market uncertainty and increasing risk, some aspects of the problem of sustainable development of the banking sector require additional scientific research. The scientific novelty is that the study puts forward a hypothesis that the forecast of the value of Sberbank’s assets for the next month can be formed based on the use of the DL deep learning model Random Forest Regression, which used monthly data on the volatility of shares of a number of commercial banks for 2024. The obtained result contains an increase in scientific knowledge and can be used as a methodological approach using artificial intelligence to assess the sustainability of commercial banks, which determines its practical significance. The purpose of the study is to obtain a forecast of the value of Sberbank’s assets for the next month, applying the developed DL model Random Forest Regression, using the identified patterns between changes in stock prices of a number of commercial banks, as well as their volatility. To achieve the goal, the following tasks were set and solved: to study the theoretical aspects of sustainable development of commercial banks, to develop a dataset of the DL-model "Random Forest", to form a forecast value of the result feature — the value of Sber’s assets for the next month. The neural network showed high forecast accuracy.