Sales forecasting at retail stores -results from Machine Learning-based software in Italy
In their participation in the SIDEA annual conference held in Marina di Orosei, Italy, the LOWINFOOD partner Roberta Pietrangelli from UNITUS presented some results about the efficacy of a newly developed machine learning technology that provides retailers with accurate forecasts of sales, a piece of research that has been done in the framework of the project.
Fresh fruits and vegetables account for 54% of the total food loss and waste in Europe, making them the most wasted food products, and also represent the main fraction, in mass, of the food waste generated at the retail level. The primary drivers behind this high wastage can be attributed to the perishable nature of fresh fruits and vegetable products and the inadequate technological equipment at retail stores to support their preservation.
Artificial intelligence (AI), has the potential to push these changes in food systems. Specifically, machine learning is widely studied in the field of food waste prevention to prevent overproduction, detect non-compliance causes and target products in the appropriate market, through forecasting, monitoring and grouping
The study contributed to determining whether the stores can improve the efficiency of their orders, thus avoiding surplus ordering, by using the forecasts as input data for decision-making. In turn, this helps understand and assess the extent to which these innovations can reduce fresh fruits and vegetables waste at the retail level. The results showed that considerable improvements are possible when considering the information that is available to food managers to place orders.Share on Facebook Share on Twitter Share on Pinterest