EEG-based age and gender recognition using tensor decomposition and speech features

Abstract

Extracting age and gender information from EEG data has not been investigated. This information is useful in building automatic systems that can classify a person into gender or age groups based on EEG characteristics of that person, index EEG data for searching, identify or verify a person, and improve performance of brain-computer interface systems. In this paper, we propose a framework based on PARAFAC and SVM that can automatically classify age and gender using EEG data. We also propose a method using N-PLS and SVM to improve the classification rate. Experimental results for the proposed method are presented. © Springer-Verlag 2013.