Abstract

Info

Title:: Automatic diagnosis of cardiomyopathy based on nonparametric correlation coefficient

Author(s):: LIN Zhuochen; ZHANG Jinxin; School of Public Health, Sun Yat-sen University, Guangzhou 510080, China

Keywords:: Keywords: electrocardiogram hypertrophic cardiomyopathy dilated cardiomyopathy Hoeffding D measure PTB dataset

Abstract:: Abstract: Based on the correlation between electrocardiogram (ECG) leads of patients with hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM), a method of automatic diagnosis of cardiomyopathy is proposed. In this study, a single heart beat segment separated from 12-Lead ECG signals is identified, and DCM and HCM segments are identified from healthy people. After calculation, 264 nonparametric correlation coefficient features are extracted from the fragments, and 12 features are obtained through variable selection. Then, these 12 features are inputted to support vector machine for modeling. 10-fold cross-validation is used to evaluate the model. The accuracy of the model is 99.88%±0.08%. With fewer features, higher speed and higher accuracy, the model is helpful to the automatic diagnosis of clinical cardiomyopathy and saves medical resources.