A Novel Features Selection Approach with Common Spatial Pattern for EEG Based Brain–Computer Interface Implementation

Abstract

This paper presents a novel method for the selection of spatial filters and features in electroencephalography (EEG) based motor imagery classification. The analyzing EEG data are divided into training and test sets. The training set is used to select appropriate spatial filters with dominant features. To accomplish such features, the EEG of training set is segmented again into two subsets termed as training subset and test subset. The features of both subsets are extracted using common spatial pattern. Then features of training subset are ranked using mutual information based approach. Besides, the features of test subset are also ranked according to the order of the training subset features. The initial classification performance using training and test subsets are obtained by using linear discriminant analysis. Then a grid search method is employed to select the effective number of spatial filter pairs as well as the discriminative features. Thus obtained spatial filter and features are used in actual classification accuracy of the test set of EEG. The experimental results show that the proposed approach yields comparatively superior classification performance compared to prevailing methods.