The aim of this paper is to demonstrate the effectiveness of utilizing neural networks in detection of cardiac arrhythmia. In this study we use a feed forward back-propagation neural network which is used for classifying an ECG signal as arrhythmic or non-arrhythmic. The classification performance is further compared to standard machine learning techniques which include SVM, Logistic Regression, KNN, Random Forest, LVQ, Naive Bayes. The binary classification accuracy achieved using neural networks is 89.11% which is significantly greater than accuracy of standard machine learning methods. © 2020, Institute of Advanced Scientific Research, Inc.. All rights reserved.