Forecasting of Tax Revenues to the Budget of the Russian Federation

The research is devoted to the analysis and improvement of the accuracy of forecasting tax revenues to the budget of the Russian Federation. The paper sets and solves the problems of comparing traditional econometric forecasting methods and modern approaches in the field of machine learning. The aim of the work is not only to improve the accuracy of tax revenue forecasts, but also to empirically confirm the possibility of using machine learning to solve econometric problems. The methodology includes data collection from official sources, data preparation, and the development and implementation of forecasting models. The empirical base consists of macroeconomic indicators and budget item data obtained from the website of the Federal Treasury of the Russian Federation. The work demonstrates that machine learning methods can be effectively used to predict tax revenues and that they can be more accurate than traditional econometric methods. The main results are an established improvement in the accuracy of forecasts and the development of a model that can be adapted for more efficient budget planning. The work also offers a methodological contribution to existing approaches to tax forecasting in the Russian Federation, expanding them by introducing new approaches to the analysis of economic data.