Research on Product Recommendation Based on Web Space-Time Customer Behavior Trajectory

Customer interest point prediction is the key to improving the accuracy of e-commerce recommendation in big data environment. The current technology mainly predicts the existing customer interest, and does not comprehensively consider the impact of customer's multiple behaviors, time sequence and time on product recommendation. A prediction method of interest points based on the spatio-temporal behavior track of customer web is studied, it is necessary to construct a client web spatio-temporal behavior hypernet model, which includes customer, time, behavior and interest point four-layer subnet. At the same time, we introduce behavior influence factors and propose a prediction algorithm of interest points based on super edge similarity to solve the problem of customers' multiple behaviors and timing time impact on recommended products, so as to improve the successful efficiency of recommendation.