Abstract: Cardiac arrhythmia is a serious and life threatening condition of abnormal heart rhythm. Cardiologists mostly rely upon Electrocardiogram (ECG) to diagnose various cardiovascular disorders. Due to various technical limitations in the visual or conventional analysis of ECG, Computer Aided Design (CAD) and analysis of ECG is introduced. Many automatic algorithms were developed for the prediction of cardiac arrhythmia from ECG signal. This paper presents a simple algorithm for the prediction of Cardiac arrhythmia from the ECG signal using Artificial Neural Network (ANN). The input to the classifier is the morphological and temporal features extracted by means of the Pan Tompkins algorithm from different ECG signals obtained from MIT-BIH arrhythmia database and PTB diagnostic ECG database. The results were evaluated in MATLAB and satisfactory results obtained with a classification accuracy of 98.8%.

Keywords: Cardiac arrhythmia, ECG, Artificial Neural Network, Confusion matrix, Classification accuracy