ICA-Based EEG Feature Analysis and Classification of Learning Styles

Abstract

A thorough investigation of the electroencephalograph (EEG) information may support an enriched awareness of the mechanism of understanding different learning styles patterns. Wavelet analysis is a powerful technique that uniquely permits the decomposition of complex information of trends, discontinuities, a repeated pattern. The purpose of such methods is to be able to assign simple segments at diverse locations and scales, to be remodelled afterward effectively. In this paper, we attempt to classify individual cognitive learning styles using artificial neural networks and unsupervised learning. First, we apply Independent component analysis (ICA) to extract relevant features (artefacts removal) of the EEG records. We analyse the ICA-based EEG channels data using inter-quartiles to show the degree of dispersion and skewness. Next, self-organising maps (SOM) are then created to characterise different cognitive learning styles from selected ICA-based channel data.