Construction Of Diabetes Nutrition Knowledge Graph

Objective:The diabetes knowledge graph including population classification,nutrition recommendation,TCM diet therapy,etc was builded by sorting out the knowledge of diabetes and its nutritional treatment which was based on the nutritional needs of people with diabetes,Diabetes nutrition knowledge,Clinical guidelines,Clinical practice,Expert experience,etc,and constructing association rules and knowledge model,to provide a reference for diabetes nutrition knowledge management and daily diet recommendation.Contents:1.Theoretical research:The research progress of medical knowledge graph and diabetes nutrition knowledge were reviewed.The key points of diabetes prevention and treatment and relevant knowledge of nutritional treatment for different types of people were summarized by systematic review of diabetes guidelines,expert consensus and textbooks.2.Design of the knowledge graph construction framework applicable to the classification of nutritional needs of different diabetic populations:Combined with the characteristics of diabetic patients,the nutritional needs of different segments of the population were analyzed,and the rules between the nutritional needs of the segment population and food were designed.3.Construction of diabetes nutrition knowledge graph:The nutrition knowledge graph architecture was designed based on the general knowledge graph architecture.4.Recommendation and application of diabetes nutrition knowledge:Personalized nutrition programs are recommended for different types of people by designing the recommended application model of diabetes nutrition knowledge.Methods:1.Literature research:By consulting Authoritative guidelines,industry standards,expert consensus and other information,to understand the knowledge graph and research development status of diabetes nutrition recommendations;To summarize the nutritional treatment of diabetes in different populations.2.Expert consultation:To solve the problems of concept determination and method selection in the design and construction of ontology and knowledge graph,as well as the key points of attention in the construction of diabetes nutrition knowledge graph by using ontology method,etc.3.Knowledge graph construction methods:A top-down approach is adopted to construct the knowledge graph.The association of nutritional treatment knowledge of diabetes was sorted out,which included the determination of knowledge source,knowledge representation,knowledge extraction,knowledge fusion and knowledge utilization.The seven-step method was used to construct the knowledge graph model layer by protege.The data layer was constructed by Neo4j.Results:1.Research on the content framework of knowledge graphBased on the top-down principle and closely combining the nutritional needs of different types of people with diabetes and its complications,the diabetes nutrition knowledge graph is constructed,which involves demand analysis,knowledge model,application model,knowledge source,knowledge analysis,knowledge representation,visual display and other links.The diabetes nutrition knowledge graph contains dietary prescriptions,knowledge rules,food lists,recommendations and contraindications,etc.,and presents the rules between nutritional needs and food of segmented populations in a graphical way.2.Research on realization of knowledge graphFirstly,the demand of diabetes nutrition knowledge graph was discussed from the dimensions of policy demand,user demand,business function and non-functional demand.Based on this,a nutritional treatment knowledge model is constructed,which involves five steps:diagnosis of diabetes,labeling of special patients,calculation of total calories and allocation of nutrients,to realize personalized dietary recommendation for users.Then,the construction process of the knowledge graph was described in detail from the aspects of knowledge sources,knowledge analysis and knowledge representation,and the association of nutritional treatment for diabetes in different populations was visualized with Protege and Neo4j tools.3.Application design and presentationFirst of all,the functions of the diabetes nutrition knowledge system which included patient data collection,food material attribute storage,nutritional recipe recommendation,food therapy formula inquiry,and data update and modification were designed from the perspectives of nutrition management,personalized service,and convenience of doctor-patient applications.Then,knowledge retrieval and application of nutrition and health education prescription were taken as examples to demonstrate the application effect of diabetes nutrition knowledge system.Conclusions:The diabetes knowledge graph was builded to discover associations among different populations,diabetes knowledge,nutrition,and diet,to provide a reference for diabetes nutrition knowledge management and daily diet recommendation.