An autometic system for ECG arrhythmias classification

An automatic system integrating the artificial intelligent methods could classify the normal sinus rhythm (NSR) and three arrhythmic types from the continuous ECG signals obtained from the MIT-BIH arrhythmia database. In this system, a support vector machine (SVM) was used to mark the heart beats of ECG with Lead II and its slope signals. An algorithm according the markers extracted segment´s waveforms of Lead II and V1 as the pattern´s features of classification. A self-constructing neural fuzzy inference network (SoNFIN) was used to classify NSR and three arrhythmic types including premature ventricular contraction (PVC), left bundle branch block (LBBB), and right bundle branch block (RBBB). The results indicated the accuracy achieved 98.9%. The accuracy of heart beat detection could be arisen to 99.3%.