A recommendation model construction method for a manufacturing service supply chain scene

Abstract

This invention discloses a method for constructing a recommendation model for manufacturing service supply chain scenarios, belonging to the field of manufacturing service supply chain technology. The method addresses the issue of insufficient spatiotemporal feature modeling in manufacturing service supply chain scenarios by constructing a knowledge graph containing spatiotemporal information. Given that existing technologies neglect business relationships between entities, this application significantly improves the accuracy and real-time performance of supplier recommendations by fusing unstructured business text with structured spatiotemporal data and utilizing a knowledge distillation mechanism to transfer the higher-order reasoning capabilities of the knowledge graph to the student model. Furthermore, it demonstrates higher accuracy and practicality in handling complex spatiotemporal dependencies and dynamic business preferences, providing strong support for supplier decisions in manufacturing enterprises.

Images