Construction Of Knowledge Base Of Diabetes Management Based On Ontology

ObjectiveBased on the management needs of diabetes mellitus at the grass-roots level and based on the clinical guidelines of diabetes mellitus,the ontology technology is used to construct diabetic ontology,comb the knowledge concepts of diabetes diagnosis,complications,treatment,and establish the semantic relationship.On this basis,design the semantic retrieval model of the diabetes ontology,in order to convert the static text form of the guide into inquiring and inferential semantic retrieval system 3 to assist the primary clinicians to make decisions better.It also provides reference for the development of diabetes decision support knowledge base system.Contents1.Relevant theoretical research on ontology technologyAccording to the ontology-related theories,analyze and compare the advantages and disadvantages of the existing ontology construction methods and tools.Then,select the ontology construction methods and tools suitable for this study based on the knowledge characteristics of diabetes guidelines.2.Construction of diabetic ontologyAccording to the "seven-step method" and the Protege(version 4.3)tool,extract concepts from the diabetes clinical guidelines to establish a diabetes knowledge model,and form an ontology base consisting of concepts,attributes,instances and related concepts.The ontology base not only contains concepts of clinical guidelines,but also includes the relationship between concepts and concepts.3.Design of diabetic semantic retrieval modelBy adding semantic reasoning to the retrieval and combining users of the ontology,analyze the function and design principles of the ontology-based semantic information retrieval model.Summary the basic idea of the ontology-based semantic retrieval model.The methodologies and techniques used in the different modules of the retrieval model are elaborated.This can provide reference for the development of the diabetes knowledge retrieval system.Methods1.Literature investigation method:through literature and related website research,learn the research and development status of digitized clinical guidelines,ontology and semantic retrieval in the field of medicine at home and abroad,comb related theoretical knowledge of ontology,and select the method and tool for constructing ontology for this study.2.Case study method:to investigate the current status and development process of diabetes guidelines at home and abroad;to investigate ontology-related theoretical knowledge,ontology construction methods and tools;and to investigate existing ontology and vocabulary in the disease domain at home and abroad.3.Expert consultation method:through consultation with diabetes clinical experts,define the scope of diabetes knowledge;through consultation with informatics experts,determine the construction method of ontology,the accuracy of the establishment of concepts and attribute relationships.4.The "seven-step method":use the "seven-step method" to construct the diabetes guideline ontology,the "seven-step method" mainly includes determining ontology's specialized fields and categories,reusing existing ontology,listing important terms,defining the levels of classes and classes,defining class attributes,defining attributes constraints and adding instances.Through the "seven-step method",tease out the knowledge structure of the diabetes guide and the relationships between the concepts.Results1.Combing diabetes guidelines and selecting construction methods and tools of ontology.On the basis of understanding the status quo of diabetes guidelines,sort out the important knowledge content of diabetes guidelines,mainly including diagnosis,complications and treatment.Analyze the advantages and disadvantages of the commonly used methods and tools of ontology,and combine the characteristics of diabetes knowledge,finally to choose the seven-step method and Protege as the construction method and tool of the ontology.2.Using the seven-step method,based on diabetes clinical guideline knowledge,MeSH vocabulary,UMLS,etc.,build the diabetes ontology knowledge base,and ultimately extract the relevant content of 8 subclasses,including diabetes type,clinical manifestations,examination,treatment,complications,co-morbidities,patient population and prevention,which covering 199 categories,21 object attributes and 54 data attributes,achieving visual presentation in Protege.3.Design of semantic retrieval model of diabetes.By analyzing the problems existing in the traditional retrieval methods and the advantages of semantic retrieval methods,design the diabetic semantic retrieval model on the basis of diabetes ontology,which mainly includes ontology management module,question word processing module,query processing module and semantic retrieval module,which can achieve better retrieval through semantic reasoning and query expansion.SummaryIn view of the problems and obstacles in the traditional clinical practice guidelines,taking diabetes as an example,use ontology technology to digitize clinical guidelines,build more comprehensive diabetes ontology,and creates inter-conceptual semantic relationships.Based on this,design a semantic ontology retrieval system model to achieve intelligent semantic retrieval of diabetes knowledge through information extraction,semantic reasoning and query expansion,to improve the accuracy and breadth of information retrieval,and provide a reference for the development of primary diabetes management system.Suggestions1.The diabetic ontology needs to be supplemented and improved with the updating of the guidelines;2.The diabetic ontology needs further empirical test and evaluation;3.The model of semantic retrieval system needs further development.