Kidney Disease Knowledge Annotation And Disease Prediction Based On Electronic Medical Records

In the era of big data,the "Internet +" medical industry has attracted the attention of researchers from all walks of life.The medical industry obtains information from electronic medical record data with the help of statistics,machine learning and artificial intelligence,and promotes the development of medical industry from disease prediction,medical risk assessment,drug discovery and other aspects.At present,following the data mining research on diabetes,heart disease,hypertension and other common diseases,the field of kidney disease has also started to cooperate with the data science,using electronic medical record data to carry out the research on the risk prediction of kidney disease.In terms of early prevention and control of diseases and prognosis of diseases,robot-assisted clinical medical work is gradually realized,and medical staff can also provide more accurate medical services for patients.Therefore,this study takes electronic medical records as the object of study,applies library information in the field of kidney disease,and conducts disease prediction research on the basis of medical knowledge annotation of electronic medical records.In this paper,based on the big data project of the hospital,the knowledge of kidney disease was annotated and disease prediction was made according to the electronic medical records stored in the hospital.Firstly,according to the literature and the experience knowledge of medical experts,the medical indexes of nephropathy were sorted out,and the indexes were mapped to the characteristics in the electronic medical record,including34 indexes including laboratory tests and clinical symptoms,and the indexes were marked and extracted in the electronic medical record.Secondly,in the task of classification and prediction of renal diseases,the XGBoost algorithm was used to construct the classification and prediction model for the auxiliary diagnosis of pediatric nephritis,to predict whether patients would develop renal damage.Through comparative experiments,the model based on the XGBoost algorithm was more effective.Thirdly,in the task of association and prediction of kidney disease symptoms,the spectral clustering algorithm was used to cluster the clinical symptoms of kidney disease,the clinical symptoms were divided into three groups of symptoms,and the clinical classification of the disease was predicted.Through the analysis of the relationship between the symptoms of kidney disease,the effectiveness of this method was verified.Finally,a big data analysis platform for kidney disease based on electronic medical records was constructed according to the knowledge annotation and the empirical results of the first two prediction experiments.This study can not only provide a new perspective for the study of electronic medical record,but also provide a reference for assisting clinical medical decision-making.