Ensemble Neural Network Algorithm for Detecting Cardiac Arrhythmia

Cardiac arrhythmias are electrical malfunctions in rhythmic beating of the heart. Sometimes, they cause life-threatening conditions. Hence, they need to be diagnosed quickly and accurately to save life and prevent further complications and effective management of the disease. In this paper, we propose an ensemble neural network algorithm to detect arrhythmia. Bagging approach with multilayer perceptron and radial basis neural networks is used to classify the standard 12-lead Electrocardiogram (ECG) recordings in the cardiac arrhythmia database available in UCI Machine Learning Repository. The classification performance of the diagnostic model was analyzed using the following performance metrics, namely precision, recall, F-measure, accuracy, mean absolute error, root mean square error, and area under the receiver-operating curve. The classifier accuracy obtained for the ensemble neural network (ENN) model is 93.9 and 94.9 % for ENN-RBFN and ENN-MLP, respectively.