Построение краткосрочных прогнозов социально-экономических показателей россии на основе последовательного применения ранжированных моделей

Abstract. The authors present a methodology for constructing short-term forecasting models using sequential application of mathematical models such as ARIMA, multiple linear regression, regression decision tree and random forest. The models in the article are ranked in order of increasing computational complexity and in order of decreasing stability and interpretability. The modeling tool is the information and analytical system "Horizon" developed by the staff of the Plekhanov Russian University of Economics.