Knowledge graph recommendation method based on comparative learning and view mining

• Main Authors: YUAN XU, QIN CHANGYUAN, WU HUINAN, GAO BUYUN, BU XIYA
• Format: Patent
• Language: Chinese; English
• Published: 08.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention provides a knowledge graph recommendation method based on comparative learning and view mining, and belongs to the technical field of recommendation systems. The method comprises the following steps: firstly, comprehensively considering and constructing four different views from different perspectives, wherein the four different views comprise a collaborative knowledge graph which is jointly constructed by a knowledge graph and a user-project interaction graph and contains global structure information; and two project-project diagrams constructed based on the user-project interaction diagram and the knowledge graph. In addition, the intersection-to-parallel ratio is applied to construct two user-user graphs, and a top-k matching mechanism is designed to avoid introduction of more noise. And then carrying out comparative learning on the four views on local and global levels, so as to mine cooperation information between users and projects, between users and between projects and global structure i