Evaluation of the effect of combining time series forecasting models

Abstract

Evaluation of the effect of combining time series forecasting models

Grebnev V. I., Nemcev M. A.,Tyukov A. P.

Incoming article date: 19.05.2020

This article investigates methods to combine multiple data-driven models for forecasting time series. This is done in order to increase the accuracy of forecasts according to the historical error signal, such that it minimises downstream costs of forecast errors.

Keywords: time series, combination of models, accuracy assessment, forecasting, forecast accuracy