Abstract
At present, there are more and more researches on knowledge graph in various fields. This paper focuses on the knowledge graph in the field of education. According to the general process of knowledge graph construction, this paper constructed electric circuits course knowledge graph. Firstly, collect original text data from two sources: print books and publicly available online resources. Then, the construction of the domain ontology of the electric circuits course is presented, which defines entities, attributes, and relations in triples. Next, the knowledge extraction part is introduced, and a comparison experiment of named entity recognition on unstructured data is conducted, comparing the recognition performance of different entity types and different models. The results show that the BERT-Bi-LSTM-CRF model has the best effect, with a precision of 91.85%. In addition, relation extraction is performed to transform unstructured data into structured triples, which are stored in MongoDB and Neo4j databases. Finally, the construction of the electric circuits course knowledge graph is completed.