MODERN ECONOMETRIC APPROACHES OF MACHINE LEARNING IN THE SIMULATION OF FACTORS INFLUENCING LOCAL BUDGET EXPENDITURES

Abstract

The article proposes to expand approaches to planning and budgeting processes. Machine learning technologies were used based on the intelligent gradient boosting algorithm of decision trees (XGBoost Tree). Reliable estimates of expenditures provided a connection between different socio-economic characteristics of regions and made it possible to predict the level of influence of factors on the expenditures of local budgets.