Product recommendation method and device, equipment and storage medium

• Main Authors: WANG JIASHUAI, LIU SHUYUE, HU YANNAN, WANG MENG
• Format: Patent
• Language: Chinese; English
• Published: 28.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

According to the product recommendation method, device and equipment and the storage medium, the social relation of customers, the historical interaction record of products purchased by the customers and the product attribute information are fused, the customer behavior knowledge graph is constructed, and the problem of data sparsity is relieved through existing semantic correlation and structural correlation. A graph attention network is adopted to relieve a coupling relationship between client intentions and product feature representation dimensions, and neighborhood feature information is aggregated under different intentions to learn feature representation of clients and products. Meanwhile, it is ensured that intention features are mutually independent, and intention redundancy does not exist. And finally, according to the decoupled feature information, capturing long-term and short-term intention preferences of the current customer, and autonomously learning an optimal recommendation strategy by using d