Abstract: Based on improving the entity alignment relationship-aware neighborhood matching（RNM） model， an entity-aligned relationship-aware neighborhood matching model combining attribute information and dual attention mechanisms was proposed. The dual attention introduced by RDGCN was used to optimize the relational structure learning ability of the original GCN. At the same time， attribute information was added， and relational structure and attribute information were combined with embedding relational aware neighborhood matching. The accuracy of alignment on three real data sets can reach 86. 91%， 87. 67% and 94. 05%， respectively， further improved compared with the benchmark model. Experimental results show that the proposed method can effectively identify the aligned entity pairs.

Key words: entity alignment, attribute information, relationship matching, knowledge fusion, knowledge graph, graph convolution network