Abstract: The scale of knowledge is growing rapidly in the big data environment, and traditional knowledge organization and services have faced the dilemma of semantic inaccuracy and untimeliness. From a knowledge fusion perspective—combining the precise semantic superiority of traditional ontology with the large-scale graph processing power and the predicate attribute expression ability of property graph—this paper presents an ontology and property graph fusion framework (OPGFF). The fusion process is divided into content layer fusion and constraint layer fusion. The result of the fusion, that is, the knowledge representation model is called knowledge big graph. In addition, this paper applies the knowledge big graph model to the ownership network in the China’s financial field and builds a financial ownership knowledge big graph. Furthermore, this paper designs and implements six consistency inference algorithms for finding contradictory data and filling in missing data in the financial ownership knowledge big graph, five of which are completely domain agnostic. The correctness and validity of the algorithms have been experimentally verified with actual data. The fusion OPGFF framework and the implementation method of the knowledge big graph could provide technical reference for big data knowledge organization and services.

Knowledge Organization, 2021, 48(1) pp. 55 – 71.