Distributional regression for demand forecasting in e-grocery

E-grocery offers customers an alternative to traditional store grocery retailing. Customers select e-grocery for convenience, making use of the home delivery at a selected time slot. In contrast retailing, in in-stock information stock keeping units (SKUs) becomes transparent customer before substantial shopping effort has been invested, thus reducing personal cost switching another supplier. As consequence, availability SKUs particularly strong impact on customer’s order decision, resulting higher strategic service level targets retailer. To account these high targets, we propose suitable model accurately predicting extreme right tail demand distribution, rather than providing point forecasts its mean. Specifically, application distributional regression methods — so-called Generalised Additive Models Location, Scale and Shape (GAMLSS) arrive cost-minimising solution according newsvendor model. benchmark models consider various as well popular from machine learning. The are evaluated case study, where compare their out-of-sample predictive performance with respect provided by retailer analysed.