Arrhythmia ECG Beats Classification Using Wavelet-Based Features and Support Vector Machine Classifier

Abstract

Abnormal behavior of heart muscles generates irregular heartbeats which are collectively known as arrhythmia. Classification of arrhythmia beats plays a prominent role in electrocardiogram (ECG) analysis. It is widely used in online and long-term patient monitoring systems. This chapter reports a classification technique to recognize normal (N) and five arrhythmia beats (i.e., left bundle branch block [LBBB], right bundle branch block [RBBB], premature ventricular contraction [V], paced [P], and atrial premature contraction [A]). The technique utilizes features of heartbeats extracted by the wavelet multi-resolution analysis. The feature vectors are used to train and test the classifier based on the support vector machine which has been emerged as a benchmark in machine learning classifier. It accomplishes the beat classification very efficiently. ECG records of the MIT-BIH arrhythmia database are utilized to acquire the different types of heartbeats. Performance of the proposed classifier outperforms the contemporary arrhythmia beats classification techniques.