[Automatic recognition of RR-intervals of sinus origin from real electrocardiographic signals].

A simple decisive rule for classifying RR-intervals of electrocardiograms by two types, sinusoidal and anomalous, is deduced using the theory of pattern recognition. Parameters of the decisive rule according to the minimum of the Buyess probability of the classification error are selected. The algorithm may be realized with microprocessors and used in cardiac rhythm monitors.