Prediction of Turnover in Gastronomy
by Means of Machine Learning Algorithms
At a glance
- Project leader : Dr. René Locher
- Project team : Loran Avci, Dr. Marcel Dettling, Dr. Christoph Hofer, Dr. Thoralf Mildenberger, Mino Müller
- Project status : completed
- Funding partner : CTI (KTI-Projekt / Projekt Nr. 26228.1 PFES-ES)
- Project partner : Prognolite GmbH
- Contact person : René Locher
Description
How many guests will visit a restaurant and at what time of the
day? Which menus will be ordered? Planning is absolutely crucial in
gastronomy but not at all easy. It must be ensured that the correct
amount of food is purchased and enough staff is present to run the
shop. The planning which has been done intuitively for the time
being can now be replaced by our machine learning algorithms.
Data from operation and cash register systems are read into a
database and used to train machine learning algorithms and to
predict turnover as well as staff needed. Relations between the
influencing factors are complex and specific to each restaurant and
yet the prognosis should be feasible with as little manual
intervention as possible. Variables to consider include time of
day, day of week, holidays, school holidays, weather conditions,
special offerings or local events etc.