Self-Organizing-Map-based Knowledge Fusion Method for Heterogeneous Ontologies in Power Grid Networks

Knowledge fusion is one of the fundamental factors of the knowledge graph technology. However in the recent big data applications with the multi-dimension and heterogeneous data structures, traditional machine learning (ML) based approaches adopted to the knowledge fusions may show limited accuracy and instantaneity. Hence, the paper proposes a self-organizing map (SOM) based on low complexity and unsupervised knowledge fusion method for ontology mapping in the heterogeneous data environment. The method targets providing efficient knowledge correlations meanwhile achieving a certain real-time requirement. It has been tested in the knowledge graph system developed by state grid Anhui electric power research institute (China). The efficiency and practicability of the proposed method is proved by comparing the performance with the traditional unsupervised ML algorithms.