Maximum Margin Clustering Method Based on Immune Evolution for Electrocardiogram Arrhythmias Diagnosis

This paper presents a novel maximum margin clustering method based on immune evolution (IEMMC) to diagnose electrocardiogram (ECG) arrhythmias. This method extends maximum margin principle from SVM to clustering and formulates the clustering model in terms of optimization problem. Then we use immune evolutionary algorithm to find out the optimal solution which has the maximum margin over all possible solutions. Five types of ECG arrhythmias obtained from MIT-BIH database are analyzed in the experiment, including normal sinus rhythm (N), premature ventricular contraction (PVC), atrial premature contraction (APC), right bundle branch block (R), and left bundle branch block (L). To assess the effect of the IEMMC method for ECG arrhythmias, attempts are then made to use three types of performance evaluation indicators, such as sensitivity, specificity and accuracy. Compared with both unsupervised and supervised leaning methods, the IEMMC algorithm reflects comprehensively superior performance in ECG arrhythmias diagnosis.