Multi-dimensional similarity entity alignment method suitable for remote sensing geographic information knowledge graph

The invention provides a multi-dimensional similarity entity alignment method suitable for a remote sensing geographic information knowledge graph. Compared with the prior art, the method for judging the similarity of the multi-dimensional entities can be provided, and equivalent entities which are easily ignored by a conventional method based on entity semantics can be mined. By means of the method, multi-dimensional similarity calculation in the aspects of semantics, attribute items and key attribute values among entities can be achieved, the equivalent entities are determined according to the calculation result, and entity alignment is carried out. Deep mining of similar entities can be met, more equivalent entities can be captured, and the accuracy of entity alignment is improved. Meanwhile, the labor cost of adding new entities into the knowledge graph is reduced, and the automation degree of entity alignment is improved. 本发明提出一种适用于遥感地理信息知识图谱的多维相似度实体对齐方法。与已有技术相比，本发明能提供多维实体相似度判定方法，能够挖掘出常规基于实体语义的方法容易忽略的等价实