An Artificial Intelligence Method for Discriminating Eye state from Quantitative EEG

Abstract

In this paper an artificial intelligence method named feed-forward back-propagation neural network (FFBPNN) is proposed for human eye state classification. The dataset consists of fourteen channels of EEG signals data in discrete form namely Quantitative EEG. The accuracy of eye state classification sometimes less due to consideration of raw EEG signal. Hence in this work, input signals are first normalized using minmax method and then given as input to the FFBPNN network. Experimental outcomes of the FFBPNN are recorded in term of ‘O's or ‘1 ‘s for two classes of eye state EEG signals. The accuracy of the proposed FFBPNN method has been measured using cross-validation process. Accuracy of the cross validation based FFBPNN is recorded up to 99.8% with 2-fold cross-validation by giving normalized data as an input to the network whereas the accuracy of the same network is recorded up to 76.5%, due to without normalized data given as a network input. Hence the proposed method gives better accuracy of the classification which will ultimately help in designing robust BCI.