In order to accomplish a well-founded production planning and control, manufacturing companies need to forecast future customer demands as precisely as possible. Different studies have shown that no single forecasting method is applicable to achieve accurate forecasts for every considered time series. Hence, this project developed an automatic forecasting system, which is able to select a suitable forecasting method, configure its parameters and compute forecasts of future values. By applying a meta-learning approach, the automatic forecasting system achieves more accurate forecasts than established industrial best practice methods.
Contact persons:
S. Oelker
Funded by:
DFG
Duration:
01.01.2013 - 30.06.2014
See project's publications
List all projects
