Classification of EEG signals by radial neuro-fuzzy system
David Coufal · 2005
Abstract: We present a hybrid system for classification of EEG signals into the three classes of mentation, relaxation and micro-sleep. The classifier is based on an neural representation of a ra-dial conjunctive fuzzy system. Conjunctive fuzzy systems are the fuzzy systems which employ fuzzy conjunctions for representation of IF-THEN structure of their rules. Radial fuzzy sys-tems have, in addition, IF parts represented by radial functions which helps to simplify their computation model. GUHA data mining method and genetic algorithms are used for learn-ing the classifier.