Classification of heart disease from ECG signals using Machine Learning

The main objective of this paper is to extract features and classify cardiac ventricular arrhythmias from ECG signals by using Wavelet Transform and machine learning algorithms. Heart-related diseases can be detected by acquiring Electrocardiogram (ECG) signal from the subject. The offline ECG data obtained from the MIT-BIH database and pre-processed by using Discrete Wavelet Transform (DWT) technique so that the low-frequency and high-frequency noises were removed. The peak amplitude and R-R interval of the ECG signal are detected and features extracted using DWT. The patient details in the database with features were combined and exported as an excel spreadsheet. This data spreadsheet is fed to the classifier using machine learning algorithms. The results were compared with different machine learning algorithms such as support vector machine (SVM) and Naïve Bayes. The linear SVM algorithm provided the highest accuracy of 99.4%.