Detection of Acute Coronary Syndrome using Electrocardiogram Signal Analysis

Acute Coronary Syndrome (ACS) is a cardiac disorder which has a major impact on deciding the mortality and morbidity rate. There is a need for early diagnosis of this heart disease using some accurate and non-invasive means. In this research, Electrocardiogram (ECG) signals sampled at 1 kHz are used to predict ACS by using signal analysis and machine learning techniques. A total of 598 subjects were engaged with this work for acquisition of data, out of which 298 were suffering from ACS and the other were normal subjects. First, the ECG signal is denoised and then preprocessed using Empirical Mode Decomposition (EMD). After the extraction of features, data is classified using the classification learner app. The system attains a maximum detection accuracy of 97% for Quadratic Support Vector Machines (SVM). The accuracy and low-cost nature of the system will be beneficial for doctors to accurately diagnose ACS.