Statistical evaluation of computational geometry and Neural network classification methods for person identification purposes via the EEG

Abstract

The Computational Geometry Algorithm (CGA) Pattern Recognition method is proposed in this work. It's development is based on known combinations of computational geometric algorithms with the aim of distinguishing between the EEGs of unrelated healthy individuals for person identification purposes. In the present study the classification results of the proposed method and those of a Radial Basis Function-(RBF) network, which were presented in our latest study, are evaluated statistically. The aim of this statistical process is to corroborate the claim that the CGA method is more efficient than the RBF network at correctly classifying EEG vectors and to verify the results of our previous studies, which showed that the EEGs of an individual carry common, specific, "genetic" features.