Construction of Neurosurgery Knowledge Graph Based on Bi-LSTM-CRF Model

Abstract

Medical knowledge is complex. The knowledge graph provides an efficient solution for integrating medical knowledge and analyzing medical data. In this paper, the neurosurgery knowledge graph is constructed. First, the ontology pattern library is constructed, and then named entity recognition is performed based on the Bi-LSTM-CRF model. The entities are extracted from the text and the entity relationships are defined. The knowledge graph of the symptom-disease-department is integrated, and finally stored in Neo4j Graph database. As the basis of information retrieval, intelligent question answering, diagnosis and treatment system, knowledge graph can effectively transfer medical knowledge and provide intelligent medical assistance for doctors and patients.