Coronary Artery Disease (CAD) Heart Beats Classification Using Recurrence Plots

Coronary Artery Disease (CAD) is the world’s leading heart disease. The main reason for CAD is fat accumulation in coronary arteries, during which the myocardium lacks the supply of blood and oxygen. Stringent blocks of the arteries may lead to heart attack or stroke. Electrocardiogram (ECG) signal is a standard diagnostic test to identify cardiac abnormalities. The nonlinearities of the ECG can be quantified using Recurrence Quantification Analysis (RQA) structures. This work discriminates Normal Sinus Rhythm (NSR) and CAD using Recurrence Plots (RPs).Further, the RQA parameters are extracted for categorizing NSR and CAD using Decision Tree (DT) and ensemble learning models. The essential (p\lt 0.05) features are fed to these ensemble models: Random and Rotation Forests (RAF and ROF), Bagging (BGG), and Adaboost (ADB) to select an appropriate classifier. Among them, ROF yielded the highest overall accuracy of 99.42%. The proposed system is cost-efficient and can be applied to discriminate the NSR and CAD heartbeats.