Switching Fuzzy Classifier for Classification of EEG Spectrograms
Abstract
The paper introduces the concept of switching fuzzy classifier. The work of the classifier is demonstrated on the classification of EEG spectrograms in different levels of vigilance of car drivers. The switching mechanism is based on the combination of C4.5 classification trees algorithm with a fuzzy classifier. The structure of the classifier arises from the concept of radial fuzzy system with nominal consequents. The system as a whole can be seen as fuzzy classifier with dynamic levels of confidence in the individual fuzzy rules in rule base.