Coronary Artery Blood Flow Prediction Based On Bayesian Network And Support Vector Machine

In recent years,the FFRCT obtained by the coronary artery computer layer photographic vascular imaging data is used to diagnose new non-invasive detection techniques of coronary functional stenosis,and become a new hot spot in clinical research.The technology combines anatomical and functional information by single check,no additional medical image collection and load drug use,but the difficulty is to simulate the real blood flow of coronary artery with the computer software according to coronary anatomical information.Therefore,coronary artery flow data is an important factor in ensuring reliable and stable operation of the simulation process when simulating calculation setting boundary conditions.The purpose of this paper is to explore factors affecting coronary blood flow in patients with coronary heart disease and construct coronary flow prediction model.The contents of this paper include the following aspects:(1)This paper summarizes the process of coronary blood flow rate and flow rate and flow research results,considering the entire cardiac cycle coronary blood circulation perfusion heart process,proposes coronary blood flow calculations considering multiple parameters such as heart rate,blood pressure,blood viscosity.ways to improve.In order to more comprehensively consider the inclusion of coronary flow calculation,systemic quantitative analysis of coronary blood flow influencing factors in META analysis,results showing that gender,erythrocyte count,red blood cell accumulation,platelet average volume,and blood in conventional detection data Uric acid,the same type of cysteine and coronary flow abnormality exhibits correlation.The results show that the above physiological parameters can be predicted as the selection of the characteristic variables as the real coronary flow rate.(2)Statistics on the data of the Harbin Hospital case in accordance with the above characteristic variable parameters,using the bayesian network to construct a coronary blood flow forecasting model.The characteristic variable parameter is used as input,and the coronary flow is used as the output.The result shows that the coronary flow prediction value calculated by the bayesian network algorithm and the measured coronary flow value average relative error and the average square root error are 16.18% and 2.074%,respectively.Preliminary completion of the prediction of personalized physiological parameters on coronary artery flow based on common suspected coronary heart disease patients.(3)Based on the same sample data,the support vector machine model optimized by the training generator is applied to coronary flow forecast.The accuracy of the bayesian network coronary prediction model is verified by the relative error of the comparison predictive model and the root mean square error.The results show that the average relative error of the support vector machine model and the average square root error are 0.565% and 0.7277%,respectively,and the corresponding error evaluation index of the bayesian network is small,and the support vector machine prediction model can be used for common physiological parameters of patients with coronary heart disease on coronary flow of patients with personalized patients,and predictive effects superior to bayes network models.The coronary flow value obtained by the multi-parameter coronary predictive model based on BN and GA-SVM this paper has better consistency with clinical measured coronary flow value.This study provides new ideas for coronary blood flow calculations,and the accuracy of coronary FFRCT has increased its role.