Research On Hypertension Drug Treatment Based On Big Data Mining

Hypertension is a common disease,and the treatment of hypertension is one of the major public health problems in the world.Currently,in the treatment of clinical hypertension,most of them are given as oral medications for the purpose of controlling blood pressure.However,because each patient has different causes of hypertension,the symptoms are different,and the combination of various drugs is very complicated.The combination of various factors makes it difficult for clinicians to give an optimal drug treatment plan for every patient.Therefore,this topic through the data mining technology to analyze the treatment of each hypertension treatment,and explore the key factors affecting the success of each treatment plan.These key factors are used as criteria to guide clinicians in prescribing patients,in order to reduce the suffering of patients,improve the quality of medical care,and save limited health resources.This topic establishes two models for data mining.One of the models is drug factor analysis model that aims to find the key factors affecting the success of each treatment plan.From the point of view of statistical visualization and machine learning algorithm,this model can obtain two feature selection results,and combine the two feature selection results to get a more complete result.In the end,every drug program can obtain a complete feature selection result,which is the key factor affecting the success of each treatment plan.The other model is the drug prediction model,five prediction algorithms for machine learning were used to model each group of drugs.Finally,the model with the highest accuracy rate was selected.The model was used to select a reasonable medication plan for patients to provide assistant diagnosis for doctors.In the drug factor analysis model,we compared the results obtained from the perspective of statistical visualization and the angle of machine learning algorithms and found that the matching degree of most drugs is as high as about 80 percent,that is,both statistical visualization and machine learning algorithms consider these factors are important.Through clinical doctors' interpretation and analysis of these results,decision information with practical significance for the selection of hypertension drugs was obtained.In the drug prediction model,we compared each prediction model and found that the model with the highest accuracy rate was random forest(75%)and the second one was gradient elevating tree(72.5%).In this data set,the ensemble learningalgorithm is superior to the ordinary algorithm in terms of actual results.This topic provides a brand-new idea for the selection of hypertension medications.After further improvement of the method,it can be extended to other chronic diseases.