Recognition of Arrhythmias Based on the Spectral Description of ECG

Recognition method of electrocardiosignals according to spectral description using multivariate discriminant analysis is the subject of this paper. Decision rules of detection of life-threatening arrhythmias, as well as recognition of cardiac irregularity, which are premonitory symptoms of dangerous states, are obtained. Results of experiments using real electrocardiogram records are provided, which demonstrated high efficiency of Fischer's discriminant application when solving the multi-class problem. Research results may be used when solving practical problems of cardiac monitoring.