Establishing A Prediction Model For Cardiovascular Disease Based On Vital Sign Parameter

Death from cardiovascular disease accounts for the first place of the total cause of death of urban and rural residents.It seriously endangers the physical and mental health of Chinese people.How to prevent cardiovascular disease is an important research content in the medical field in China.In recent years,patients suffering from cardiovascular disease have gradually increased,representing an increasing proportion of the causes of death.Therefore,we should not only do medical treatment,but also do timely early prediction.With the development of computer technology and big data,people began to use mathematical models in disease research to study the characteristics of the disease through quantitative analysis.In recent years,machine learning methods have become a research topic in disease prediction and clinical treatment,and this paper applies machine learning to the early prevention of cardiovascular disease.Focus on the cardiovascular disease model based on neural network algorithms and its optimization methods.The main research methods of this paper are as follows:(1)Considering the imbalance,redundancy and high-dimensional characteristics of the disease data,we study the data preprocessing,dimension reduction and unbalanced data processing methods,and optimize the traditional machine learning algorithm.(2)Using the neural network algorithm to establish a prediction model of the number of cardiovascular diseases.First,it is necessary to determine the number of input layer nodes,the number of nodes of the implied layer and the output layer nodes of the neural network,write a complete neural network cardiovascular disease prediction program under the Python platform,and evaluate the prediction performance of the model by calculating the average absolute percentage error,acceptance and accuracy of the prediction model.On this basis,the physical examination data of a Grade A hospital was selected,with 21 vital signs as influence factors,whether cardiovascular disease as the output variable,and the first 80% of the original data were selected as training samples,a total of 1524 samples and the remaining 274 data were used as test samples.The training data sample and test data samples consists of two sets of input matrices and output matrices,one used to train the network,while the other set of input and output matrix tests the trained network to build a second classification model.Experimental results show that the correlation analysis of five vital signs(age,systolic blood pressure,high-density lipoprotein cholesterol,triglycerides,creatinine)as a key variable for cardiovascular disease prediction,the accuracy of the disease diagnosis results is 80%,indicating that the model to effectively judge cardiovascular disease signs.