Machine Learning to reduce food waste in public meal organizations

Important results from the project

The project´s objective has been to investigate the potential of a full-scale forecasting tool, matomaticML, which through machine learning helps the end user to reduce food waste. The tool has the potential to give a forecast in public catering organizations that is 2-3% accurate compared to cooking for the number of enrolled students, which then in the worst case gives an error of 20-40%. The market potential for the tool was investigated in the spring 2020 in connection with Covid19, as interest in being able to plan meals increased due to a large drop in the number of students present.

Expected long term effects

Making a forecast of the number of people who are expected to come to a meal has proved to be relatively good, the obstacle that remains for the tool to have full effect is to make a forecast that the staff in the kitchen can trust. This is about balancing supply and demand, but at the same time minimizing food waste. With Covid19, the project has not had the opportunity to measure the effect of using the tool and its potential to reduce food waste, this is something that will be investigated further together with established project partners in the future.

Approach and implementation

Having continuously had access to the business where the zero series is to be tested has been a valuable resource. What has shown with this proximity is that it takes time for solutions to gain the trust of end users and something that will need to be evaluated in the future to be able to take the next step in a market upscaling that is then aimed more at the private sector.