A radial basis function neural network predicted eight cardiac arrhythmias from heart rate time series with an overall accuracy of 96.33%.
Does a radial basis function neural network accurately predict cardiac arrhythmias based on heart rate time series?
A radial basis function neural network can predict eight different cardiac arrhythmias from heart rate time series with high accuracy (96.33%).
This paper proposes the system to predict eight cardiac arrhythmias using the radial basis function neural network (RBFN). In our study of neural network for heart rate time series, the prediction of Left bundle branch block (LBBB), Atrial fibrillation (AFIB), Normal Sinus Rhythm (NSR), Right bundle branch block (RBBB), Sinus bradycardia (SBR), Atrial flutter (AFL), Premature Ventricular Contraction (PVC), and Second degree block (BII) is done using proposed algorithm. The heart rate time series are obtained from MIT-BIH arrhythmia database. The linear and nonlinear features are detected from heart rate time series of each arrhythmia. The 70% of each datasets of features are used to train RBFN and remaining 30% of the datasets of features are used to predict eight cardiac diseases. This approach gives overall prediction accuracy of 96.33% as compared to the methods reported in existing literature.
Kelwade et al. (2016) studied Cardiac arrhythmias. Radial basis function neural network (RBFN) vs. Methods reported in existing literature was evaluated on Overall prediction accuracy. A radial basis function neural network predicted eight cardiac arrhythmias from heart rate time series with an overall accuracy of 96.33%.