Optimasi Fuzzy Inference System pada Sistem Deteksi Fibrilasi Atrium dengan Fitur Elektrokardiogram Kemas Farosi, Nuryani Nuryani, Darmanto

Abstract

A study about detection of atrial fibrillation using electrocardiogram and fuzzy inference system (FIS) has been succesfully conducted. Statistical features of RR interval is used and they are mean and standard deviation of RR interval. FIS architecture is optimized by variation number of features, FIS membership function and defuzzification methods. Based on experimental result, the best FIS architecture in FA detection is which uses both statistical features (RRave and RRstd) at once, with gaussian membership functions and mean of maxima deffuzification method. The best architecture performances are 81,55%, 82,12% and 81,93% in terms of sensitivity, specificity and accuracy, respectively.