Research on Intelligent Organization and Application of Multi-source Heterogeneous Knowledge Resources for Energy Internet.
Yuxuan Wang,Liqun Luo,Guangjian Li
DOI: https://doi.org/10.1162/dint_a_00158
2022-01-01
Data Intelligence
Abstract:ABSTRACTTo improve the informationization and intelligence of the energy Internet industry and enhance the capability of knowledge services, it is necessary to organize the energy Internet body of knowledge from existing knowledge resources of the State Grid, which have the characteristics of large scale, multiple sources, and heterogeneity. At the same time, the business fields of State Grid cover a wide range. There are many sub-fields under each business field, and the relationship between fields is diverse and complex. The key to establishing the energy Internet body of knowledge is how to fuse the heterogeneous knowledge resources from multiple sources, extract the knowledge contents from them, and organize the different relationships. This paper considers transforming the original knowledge resources of State Grid into a unified and well-organized knowledge system described in OWL language to meet the requirements of heterogeneous resource integration, multi-source resource organization, and knowledge service provision. For the State Grid knowledge resources mainly in XML format, this paper proposes a Knowledge Automatic Fusion and Organization idea and method based on XSD Directed Graph. According to the method, the XML corresponding XSD documents are transformed into a directed graph in the first stage during which the graph neural network detects hidden knowledge inside the structure to add semantic information to the graph.In the second stage, for other structured knowledge resources (e.g., databases, spreadsheets), the knowledge contents and the relationships are analyzed manually to establish the mappings from structured resources to graph structures, using which the original knowledge resources are transformed into graph structures, and merged with the directed graphs obtained in the first stage to achieve the fusion of heterogeneous knowledge resources. And expert knowledge is introduced for heterogeneous knowledge fusion to further extend the directed graph. And in the third stage, the expanded directed graph is converted to the body of knowledge in the form of OWL. This paper takes the knowledge resources in the field of human resources of the State Grid as an example, to establish the ontology of the human resources training field in a unified manner, initially demonstrating the effectiveness of the proposed method.