The Use of the Aggregator for Choosing the Method of Forecasting Time Series

In this paper, we propose an approach to optimizing the solution of the problem of forecasting time series. Optimization of the solution consists in using an aggregator to select the forecasting methods that will give the best forecast. When choosing methods, the time series features are taken into account. The aggregator uses methods of machine learning to select methods. The aggregator is built into the information system “Combination of fuzzy and exponential models”.