Literature Review of Feature Extraction Methods for Classification of EEG Signals

In this work, we have documented and compare various feature extraction methods for classification using EEG signal. This paper contains a comparative study of data reduction methods which enhances the classification accuracy. Deep study of decomposition of signals into the frequency sub bands by wavelet method (DWT) and a set of statistical features that were extracted from the EEG signals to represent the distribution of wavelet coefficients is explained. Data dimension methods like ICA, PCA and LDA are used for the reduction of dimension of data and signal vectors which can be converted to features vectors and after data reduction by suitable selection method are fed to the classifiers and the performance and accuracy of classifiers are compare in terms of accuracy to show the excellent classification process.