USING MACHINE LEARNING ALGORITHMS FOR FORECASTING MACROECONOMIC INDICATORS

Improving the accuracy and speed of macroeconomic indicator forecasting is of particular importance in the context of economic instability and global transformations. This article explores the potential of using machine learning algorithms to build adaptive models capable of predicting key macroeconomic parameters with high precision. A comprehensive approach is developed and tested, including the preparation of economic time series, construction of regression and recurrent models, and comparative analysis of their performance across multiple metrics. Examples are provided of models applied to short- and mediumterm forecasting of GDP, inflation, unemployment, and interest rates. Special attention is given to model interpretability, anomaly resilience, and applicability in state and investment policy contexts. The results confirm the high informativeness and practical value of machine learning algorithms in macroeconomic analysis.