Research On Hypertension During Pregnancy Based On Data Mining

Pregnancy-induced hypertension and its complications are the second major influencing factors for maternal mortality,posing serious threats to pregnant women and newborns.The disease is more prominent than hypertension and proteinuria after 20 weeks of gestation,and damages other organs and systems in the body,which not only affects the growth and development of the mother,but also causes maternal and child death,which leads to maternal and fetal death.One of the main reasons.At present,there is no research in the industry to determine its pathogenesis and influencing factors.In this context,this paper uses data mining methods to study hypertension during pregnancy.The content of the paper can be roughly divided into the following parts:(1)Research on pregnancy-induced hypertension based on random forest and xgboostThe first part first pre-processes the data of pregnancy hypertension,including data cleaning,de-duplication,processing missing data and attribute specifications,and finally data transformation and modeling analysis.This chapter uses random forest and Xgboost machine learning models to model and analyze the characteristics of pregnancy hypertension.The experiment found that the characteristic scores of blood pressure and patient height and body mass index are larger than those of calcium,sodium,red blood cell and hemoglobin,which play a significant role in disease decision-making.We found that the random forest model has an accuracy rate of 82.5%,which is about 3 percentage points higher than XgBoost,but XgBoost training is faster than random forests.In summary,random forests are suitable for predicting hypertension during pregnancy.(2)Classification and prediction of pregnancy-induced hypertension based on fusion modelIn view of the fact that a single basic model cannot extract all the potential laws of data,in order to exploit the advantages of various models and improve the accuracy of classification prediction,this chapter proposes a fusion model based on random forest and Xgboost,using Stacking in integrated learning that is a two-layer structure model.Through experiments,the accuracy of the fusion model is about 83.68%,which is about 1 percentage points higher than the accuracy of the single model prediction.The fusion model has better resolving power for the data and better model performance,which can be applied to the study of medical diseases.(3)Pregnancy data acquisition system based on java webThis part realizes a pregnancy data acquisition and prediction system by means of java web related technology.The Pregnancy Data Acquisition and Prediction System uses JavaScript,JQuery and Bootstrap technologies currently widely used in the Internet industry,as well as Spring,SpringMVC and Mybatis open source frameworks,combined with mature MySQL database technology.The system provides login registration,data entry query modification and data statistics functions,and develops administrator rights to manage the entire system data in a unified manner.The fusion model algorithm is integrated in the administrator interface to realize the prediction of hypertensive diseases during pregnancy.