NovEA: A Novel Model of Entity Alignment Using Attribute Triples and Relation Triples.

Entity alignment is the foundation of knowledge fusion, which can find the alignment relationships between entities in heterogeneous knowledge graphs. However, traditional methods rely on external information and need to construct data features manually. Meanwhile, embedding models do not fully utilize the pertinent information of attributes in the knowledge graphs, which limit the role of attribute information in entity alignment. Considering the shortcomings of existing methods, this paper proposes a novel model named NovEA that using attribute triples and entity triples in the knowledge graphs to complete the entity alignment task together. Besides, for attribute triples, we propose a method that can automatically generate the optimal attribute according to the data characteristics to constrain the result of attribute triples alignment and improve the accuracy of entities in alignment. Finally, we use a binary regression method to measure the similarity of the combination results of structure and attribute. Our research on real datasets shows that the NovEA model has a significant improvement in entity alignment compared with the most advanced methods.