Planning and relaxed state EEG signal classification using complex valued neural classifier for brain computer interface

Most of the Brain Computer Interface (BCI) techniques use EEG signals as a main source. Any BCI system consists of three modules and they are signal recorder, signal preprocessor and classifier. Development /Selection of efficient classifiers are a challenging task in this domain. The key work addressed in this paper is the classification of EEG signals measured under planning and relaxed state using advanced machine learning classifiers. Planning relax dataset is a benchmark data and it is obtained from UCI (University of California Irvine) machine learning repository. FC-FLC is a recently developed fast learning complex valued classifier and it is used for the EEG signal classification task. Complex valued classifier (FC-FLC) performs better than all the real valued classifiers as well as few fuzzy classifiers taken for comparison from the literature. The improvement is due to the use of Gd (gudermannian) activation function in the hidden layer of the network and the tuning free algorithm.