Construction Of Diabetes Knowledge Graph And Its Application Research

With the aging of the population and changes in lifestyles,diabetes has become a disease that seriously affects public health in my country.As a chronic disease,diabetes has the characteristics of low awareness and low control rates in my country,which has brought great harm to patients and the country.burden.The patient’s self-management ability is one of the key factors for the effective control of diabetes,which depends on the patient’s mastery of diabetes knowledge and skills.Knowledge graph can connect knowledge through semantics to form a semantic network.In this paper,a diabetes knowledge graph is constructed,and a diabetes knowledge graph based on the needs of patients is constructed through the steps of knowledge collection,knowledge sorting,and graph construction.In the construction process,the examples of diabetes problems are classified and sorted,which reflects the actual needs of patients and lays a foundation for later application.In the later stage of this study,a diabetes intelligent question answering system based on knowledge graph was constructed,aiming to explore the practical application of diabetes knowledge graph and provide consultation and answering services for diabetes-related people in daily self-management,which can not only effectively improve users’ diabetes self-management ability,It can also provide auxiliary medical services for primary medical and health workers,and provide a new way for diabetes prevention and treatment.The main research contents of the article are as follows:(1)Standardization of diabetes data.Knowledge collation is the cornerstone of all work.The knowledge sources of this study are mainly the following three aspects: first,relevant information released by authoritative institutions;second,relevant literature;.The obtained question examples and diabetes knowledge were sorted out respectively,and 43 types of questions and corresponding standard answers were summarized,as well as the knowledge of diabetes related to the questions.(2)Construction of diabetes knowledge graph.This paper adopts a top-down approach to construct a diabetes knowledge graph.The schema layer builds a semantic network with 40 semantic types and 12 semantic relations.In the construction of the data layer,the question type is used as the entity,and the standard answer is used as the attribute value,which provides the basis for the construction of the question answering part of the intelligent question answering system;The knowledge system is formed through the connection of semantic relations in the problem types,the construction of the schema layer and the data layer is completed by the combination of manual annotation and computer processing,and knowledge modeling is carried out using Protégé 5.0.0,and finally stored in the Neo4 j graph database.(3)Application of diabetes knowledge graph in intelligent question answering system.Apply the constructed diabetes knowledge graph to the construction of diabetes intelligent question answering system,the overall system construction includes two parts: problem understanding and problem solving.Among them,question understanding includes four modules,namely: LTP analysis,semantic keyword and main relation word extraction,question classification,and entity linking;question answering includes three modules,namely: query generation,graph query and answer generation.The innovations of this research are as follows:(1)The construction of this research is based on the actual needs of users’ daily questions.In the knowledge sorting stage,through the analysis of the examples of questions raised by users,according to the different knowledge points contained in the questions The questions are classified and standard answers for different types of questions are constructed,which provides the realization basis for the question answering part of the intelligent question answering system.(2)The construction of knowledge graph connects knowledge through semantics.In this study,examples of diabetes problems and corresponding diabetes knowledge are connected through semantics,so that the organization mode of diabetes knowledge is no longer organized according to traditional knowledge construction,but according to The user’s questioning needs are sorted into knowledge,which is closer to the specific needs of patients in use.