Collaborative filtering method, device, server and storage media combined with time factor

摘要

The collaborative filtering method combined with a time factor includes: building an exponential smoothing model; Acquiring a suggested period for the exponential smoothing model, the period comprising a plurality of time periods; Acquiring a user preference value of the user identifier for a specific product for the plurality of user identifiers and a plurality of time periods; Performing an iterative calculation of the user preference value using the exponential smoothing model and obtaining a smoothing result corresponding to the time period; Generating a sparse matrix using the user identifier and the smoothing result corresponding to the time period, wherein the sparse matrix includes a plurality of user preferences to be predicted; Obtaining a collaborative filtering model and inputting the smoothing result corresponding to the time period into the collaborative filtering model; And calculating and obtaining predicted values of the plurality of user preferences to be predicted from the sparse matrix by learning through the collaborative filtering model.