Heterogeneous graph embedding aims to learn a low-dimensional vector representation for each node of different types in the graph.This representation can be widely applied in various network analysis tasks such as node classification,link prediction,etc.Ex-isting methods often face challenges such as losing some key,fine-grained relational information and insufficiently handling high-order neighbor nodes when processing heterogeneous graph embeddings.To address these issues,this paper proposes an Attention-aware Multi-relational heterogeneous graph embedding method.Specifically,the method guides node aggregation by adding a decay-based high-order co-neighbor similarity matrix to the heterogeneous graph neural network and learns the importance between nodes.This ma-trix uses a decay exponent to ensure that the closer the node is to the current node in terms of hops,the greater its contribution to the current node.In this way,the model can effectively learn the importance relationships between nodes in the graph and capture inter-node relationships.Compared to existing graph adaptive attention models,this model offers more intuitive interpretability.Extensive experiments demonstrate that the proposed model outperforms the state-of-the-art baseline models.
万方数据
维普
