For testing the efficiency of arrhythmia classification under different RRI attributes, we take the ECG records (R207, R124 and others) from MIT-BIH database as training data, and adopt RBF SVM as the default classification method. Here, in our experiments, RBF SVM is with the optimal parameters (sigma and C) [42]. In practical, the system can be set with the optimal RBF parameters for personalized service. Table 2 summarizes the arrhythmia notations and counts in R207 and R124. And shown in Table 3, we denote RRI_1 and RRI_2 by 2-RRI, denote from RRI_3 to RRI+6 by 9-RRI, and denote from RRI_24 to RRI+48 by 72-RRI.