SG-CIM Mapping Entity Relationship Inference and Verification Techniques Based on Graph Representation Learning Technology
Yu Huang, Wenhui Hu, Xueyang Liu · 2021
According to the problems of entity intelligent matching and relationship prediction of power grid model, this paper adopts the entity-relation triplet automatic reasoning method to predict and correct the relation between the given entities. Firstly, the relationship in the SG-CIM model is analyzed, aggregated, normalized, and the data set of triples is constructed. Then, association rule mining and knowledge reasoning model TRANSE are adopted to model the relationship on the low-dimensional entity representation space, which is used as the translation between embedded Spaces. Thus, it is easy to capture the semantic relationship between entity pairs and improve the intelligent reasoning ability of the model.