REWE Informations-Systeme GmbH, Funding period 2003 - 2004

Automated supply of food retailers has been implemented by REWE, the
leading german company for food retailing and wholesaling, already in the
mid-nineties. Meanwhile sales forecasting is based on the daily sales data of
every single item in every single store, which is automatically provided by the tills.
For a single store the number of transferred daily sales data lies in the range
of several thousands. Consequently, there is a strong need for forecasting
methods that are able to simultaneously processing large groups of items in
reasonable time with satisfactory results. Solved problems include:

• Classification and analysis of correlation effects
• Improvement of forecasting through using correlation effects
• Automated detection of outsold items
• Correction of outliers
• Modelling the effects of (public) holidays