Classification of paroxysmal atrial fibrillation patients with SD1, SD2 and CCM features

Advances in medical technology allowed the possibility of avoiding arrhythmias with early diagnosis. Then, it is important to classify ill and healthy people correctly. In this work, the effect of Poincare plot measures (SD1 and SD2) and complex correlation measure (CCM) on the classification of subjects with and without paroxysmal atrial fibrillation (PAF) is examined. SD1, SD2 and CCM features, which were obtained from five minute RR series, were classified with k nearest neighbor classifier. The results show that using these three parameters together gives more successful results.