Research On The Construction Of Growth And Development Knowledge Graph Based On Multi-modal Data

The wish of the family is the children’s healthy growth,one of the most concerned issues of the country,society and parents is the growth and development of children.Children’s growth and development problems focused on children’s growth and development of all kinds of common disease,according to the clinical manifestations,can be subdivided into obvious symptoms of disease,such as dwarfism,gigantism,etc.,and the early symptoms of these diseases is less obvious and subtle,which involves many factors and very complicated,leading to medical workers,especially general practitioners at the basic level,to diagnose the condition.As a kind of knowledge management technology,knowledge graph stores the concepts of things in the real world and the relations between different things,which opens up a new technical idea for improving the status quo of "knowledge island" in the field of medicine and achieving the integration of knowledge resources.If the knowledge of children’s growth and development problems is analyzed,sorted out and summarized,it will have great practical meaning of the disease field’s diagnose and treat.Therefore,this paper aims to build a growth and development knowledge graph,in order to promote the efficiency of clinical medicines’ diagnose and treat,reduce the cost of doctors’ practice and patients’ medical expenses,and ultimately benefit patients.In this article,aims to research the construction of growth and development knowledge graph,mainly including the following contents:(1)In view of the key problems in the diagnosis and treatment of children’s growth and development diseases,according to the "Implementation Plan for Standardized Growth and Development Diagnosis and Treatment Center" diagnosis and treatment guidelines and the clinical practice recommendations of doctors and experts as knowledge standards,the conceptual model layer is designed,including logical paths such as medical terms,concepts and relationships;The electronic medical records of patients with growth and development diseases and the drug instructions obtained by crawler technology were supplemented as the data source of this research topic.(2)A growth and development knowledge map way using multiple knowledge languages and based on multi-modal is proposed.For text data,manual and named entity recognition techniques were used to extract entity sets according to different structure types,and entity layer was filled for conceptual pattern definition of knowledge graph.For the image data,use digital image technology to deal with the facial pictures of patients in the electronic medical record,extract the relevant facial features of various diseases,further enrich the data and knowledge sources of the knowledge map,and enhance its functions.(3)To research and develop a growth and development learning question-answering system,to demonstrate its processing process on the data of real children with growth and development diseases,and to provide pathological knowledge inquiry services for patients in the form of knowledge question answering.