Digital Deep Learning Model for Forecasting the Key Rate of The Central Bank of The Russian Federation in Volatility

The article considers theoretical issues of forecasting the key rate of the Central Bank in order to ensure the stability of the Russian banking system in conditions of volatility. The relevance of the study is that in modern conditions, artificial intelligence systems are widely used, and approaches are explored that allow fi lling the gap in theoretical issues related to modeling the mechanism for changing the key rate to ensure the stability of the banking system. The scientific novelty lies in the fact that the study puts forward and proves the hypothesis that with the help of the artifi cial intelligence system "DL-model "Random Forest" an accurate forecast of the key bank rate can be obtained. The practical signifi cance is that the results can be used in practice to provide support for decision-making regarding the level of the key rate to ensure infl ation targeting and ensure the stability of the Russian banking system. The model dataset size was 8 columns of 103 rows, it included parameters refl ecting the development of the banking system for the period from January 2016 to July 2024. The model showed good values, so the mean absolute error (Mean Absolute Error) was 0.53571; the mean squared error (Mean Squared Error) was 1.28273 and the squared root mean squared error (Root Mean Squared Error) was 1.13258. The forecast value of the key rate was 19.07834779 %, which almost coincided with the level of the rate set by the Central Bank on September 13, 2024 (19 %).