Classification of Arrhythmia by Analysing Ecg Features

Abstract

The rate of heart disease is increasing at an exponential rate. Modern-day lifestyle and our ignorance of health have put the most important organ of our body Heart at great risk. Currently, India is witnessing a large number of young people suffering from Heart Disease. Prediction of heart disease is a difficult and risky task. Since it is directly dependent on people's health, accuracy is a major factor. Today, many hospitals maintain records of their patient's reports. This kind of framework generates an enormous amount of information that can be visualized. Moreover, this information can be used in the decision-making about Heart disease. The variation in heart rate from the normal rate indicates the abnormal behavior of the heart which is a symptom of heart arrhythmia. The proposed work is to develop a computerized diagnosis method for the detection and classification of heart arrhythmia by examining the ECG signal. The ECG signal analysis is a very simple, reproducible, and inexpensive method. The study concerns four types of arrhythmias databases from MIT-BIH data namely: Standard Arrhythmia ECG Data, CU Ventricular Tachyarrhythmia Database, Supraventricular Arrhythmia Database, and Ventricular Tachyarrhythmia Database. The performance is evaluated for different machine learning classifiers.