User demand forecasting methods, devices, storage media, products and electronic equipment

Abstract

This application provides a user demand prediction method, apparatus, storage medium, product, and electronic device. The method, applied in the field of computer technology, includes: inputting sample user demand data into an initial user demand prediction model to generate customized user representations for each day within a prediction period; using the initial user demand prediction model to generate a predicted user demand probability distribution within the prediction period based on the customized user representations, wherein the predicted user demand probability distribution includes the demand probability value of the sample users generating demand each day within the prediction period; adjusting the parameters of the initial user demand prediction model based on the actual user demand probability distribution within the prediction period and the predicted user demand probability distribution, until the initial user demand prediction model completes model training to obtain a user demand prediction model; and performing demand prediction processing for target users based on the user demand prediction model.

Images