Graph neural network recommendation method based on fusion enhancement of knowledge graph and social graph

A graph neural network recommendation method based on knowledge graph and social graph fusion enhancement comprises the following steps: 1) an initial embedding layer is dedicated to coding user, project and entity nodes, and converts the attributes into vector representation in a high-dimensional space; 2) fusing features from different sources to enhance the characterization capability of the model; 3) realizing effective propagation and iterative updating of information through a multi-layer structure of the graph neural network; 4) integrating and splicing the node representations after multiple rounds of iteration updating to form comprehensive feature representations; 5) dynamically adjusting a weight ratio between the social graph and the knowledge graph to ensure that the model can be flexibly adjusted according to different scenes so as to achieve an optimal recommendation effect; and 6) using the dot product of the user and project embedding as a final score, using a BPR loss function, matrix decomp