Research on Question Answering over Knowledge Graph of Chronic Diseases

Abstract

The knowledge graph is a kind of semantic knowledge base, which can efficiently manage massive knowledge. Question answering over knowledge graph is one of the promising approaches to obtaining large information from the databases, which is applied to reply to natural language questions using structured relationship information between entities stored in knowledge graphs. Chronic disease care can be lifelong, scientific protection can have a positive effect on the recovery of patients, and lessen the incidence of complications. Therefore, this paper uses the knowledge graph to manage medical information about chronic diseases and provides users with consulting services on health issues to assist in diagnosing and treating diseases. In the paper, data is extracted from the healthcare website and stored in the graph structure database, Neo4j, after a series of data processing. We get upper and lower relationships from the data and use the extraction method, which is proved to be effective, for the domain knowledge graph construction. The platform based on high-quality knowledge, which is provided by the knowledge graph, can effectively identify user intentions and give accurate results. The research is aimed at chronic diseases and can supply references for identification, treatment, and patient self-care.