An approach to methods for processing uncertain factors in short-term macroeconomic forecasting

Abstract

The problem of how to get a final forecast by means of integrating both the forecasts of several models and the influence of large quantities of random factors that cannot be included in the forecast model is addressed. A method of solving this problem by using a structural model for processing the mathematics or systems models and semistructural methods for processing the influence of all kinds of random factors is presented. Practice has demonstrated that the results are much better than that of single forecasting model processing.<>