Research On Cardiovascular And Cerebrovascular Disease Prediction Based On Gate Recurrent Unit

In order to cope with the current situation of increasing morbidity of cardiovascular and cerebrovascular diseases in China,the scientific prevention of cardiovascular and cerebrovascular diseases has become a key project of China's public medical and health services.By predicting the future morbid state of cardiovascular and cerebrovascular diseases,screening out high-risk groups and carrying out scientific medical interventions on them can effectively prevent the occurrence of cardiovascular and cerebrovascular diseases.However,the current cardiovascular and cerebrovascular disease prediction methods generally predict people's future morbid state based only on people's current health conditions,while ignoring the close relationship between the historical changes in people's health conditions and the morbid state of cardiovascular and cerebrovascular diseases.In response to this problem,this paper proposes a method for predicting the morbid state of cardiovascular and cerebrovascular diseases based on the gate recurrent unit(GRU).GRU is a neural network model with time-series characteristics,which can fully mine time-series data such as health records.In this paper,we first determine the predictive indicators of the model by studying the main risk factors of cardiovascular and cerebrovascular diseases,then,collect the data and preprocess the data to obtain a usable data set,and use the GRU to establish a cardiovascular and cerebrovascular disease prediction model.On this basis,we designed and implemented a cardiovascular and cerebrovascular disease prediction system,and can predict the future morbid state of cardiovascular and cerebrovascular diseases through this system,and then realize personal health management and disease prevention.The main research contents of this article are as follows:(1)Collection and processing of cardiovascular and cerebrovascular diseases data.Through the research on the existing cardiovascular and cerebrovascular disease risk scoring system and related literature,the main risk factors of cardiovascular and cerebrovascular diseases were screened and used as the predictive indicators of the prediction model.Then,we extracted data from the database,preprocessed the extracted data,and obtained a data set that can be used to build a prediction model;(2)Construction of the prediction model of cardiovascular and cerebrovascular diseases.Based on the GRU,a cardiovascular and cerebrovascular disease morbid state prediction model was established.During the process of model construction,grid search and cross-validation were used to select the hyperparameters of the model,and the problem of overfitting during model training was solved.Finally,the trained model is tested,and the test results show that the prediction performance of the GRU model is higher than that of the traditional machine learning model for health record data with time-series characteristics.Using this model can accurately predict the future morbid state of cardiovascular and cerebrovascular diseases;(3)Design and implementation of the prediction system of cardiovascular and cerebrovascular disease.Based on the prediction model of cardiovascular and cerebrovascular diseases,according to the actual needs of a hospital and the service community,through the object-oriented design method,completed the design of the overall functional structure of the cardiovascular and cerebrovascular disease prediction system,and implemented the main functional module of system by using Java.In this paper,a series of researches on the prediction method of cardiovascular and cerebrovascular diseases based on GRU are carried out,the purpose is to effectively predict and intervene cardiovascular and cerebrovascular diseases before they occur,so as to provide a valuable technical approach to improve the quality of prevention work of cardiovascular and cerebrovascular diseases and reduce the morbidity of cardiovascular and cerebrovascular diseases.