A Study On The Computer Aided Diagnosis Of Chest Pain Triad

BackgroundChest pain triad has similar symptoms of chest pain in the clinic,all of them are acute and severe,and the rate of misdiagnosis is high.In recent years,data classification technology has been widely used in the diagnosis and analysis of diseases,and the effect is remarkable;classification methods mainly include decision trees,artificial neural networks,support vector machines,etc.among them,support vector machines have excellent performance in processing high-dimensional small sample data sets.However,the traditional classification method is not effective in solving the problem of imbalanced data classification.Therefore,the computer aided diagnosis of chest pain triad based on mixed feature selection can help doctors to quickly and accurately diagnose the results and formulate reasonable treatment strategies.ObjectivesBased on the medical history and examination report of patients with aortic dissection,pulmonary embolism and acute myocardial infarction,CS-SVM and mixed feature selection method were applied to the classified diagnosis of chest pain triad,thereby further improving the classification accuracy rate and reducing the time complexity,and providing reference guide for the clinical diagnosis and treatment of the chest pain triad.MethodsOn the basis of support vector machine and feature selection,CS-SVM and mixed feature selection are applied to the classification of the chest pain triad,respectively.Among them,CS-SVM is applied to the classification and recognition of the chest pain triad,and the mixed feature selection is applied to the dimension reduction of the chest pain triad data set,and the classification accuracy and time complexity were used to evaluate its effect.ResultsIn this study,the classification of chest pain triad syndrome based on CS-SVM is studied,and the experiments are carried out based on balanced data set and non-equilibrium data set,respectively.Based on the balanced data set,the comparative experiment of classical SVM and CS-SVM is used to verify the advantages of CS-SVM.Based on the imbalanced data set,the classical SVM,PSO-SVM,GA-SVM and CS-SVM comparative experiments are used to verify the classification effect of CS-SVM,and the five-fold crossover and random sampling were used to ensure the rationality and scientificity of the experiment.The experimental results show that the classification accuracy of CS-SVM is improved to some extent compared with other classification algorithms,and the accuracy of the five-fold crossover is stable and the time complexity is low.Aiming at the dimension reduction of chest pain triad data set applied to mixed feature selection,experiments are carried out based on open data set and data set in this study.That is,the classification accuracy of CS-SVM is compared with that of mixed feature selection.The experimental results show that based on the open data set,the classification accuracy of mixed feature selection is greatly improved and the time complexity is greatly reduced.Based on the data set of this study,the five-fold crossover accuracy has been improved and the time complexity has been greatly reduced.The experimental results show that the hybrid feature selection method can obtain fewer features related to classification and obtain better classification results.