A fuzzy classifier based detection for epileptic seizure signals

The present work aims to propose a system for classifying the Electroencephalogram (EEG) signals by using Fuzzy inference system. First, we denoise the signal by Short Time Fourier Transform (STFT). Then, we decompose EEG signals into six subbands, namely D1 to D5 and A5 by using Wavelet Transform (WT) for extracting features. Finally, we classify 4 EEG classes; which are Normal, Interictal, Focal and Non-focal. This classification was conducted by the Fuzzy classifier which is consisting of Fuzzifier, Inference Queue, Defuzzification Unit, and a knowledge Based unit. While The best available detection system achieved about 96% detection accuracy for data-set size 500, the proposed system achieved the same detection accuracy for data-set size 1050 (i.e. More than double the size of the previous dataset).