THE EFFECT OF USING HAMMING WINDOW AND LINEAR PREDICTIVE CODING MODEL IN EEG-P300 SIGNALS CLASSIFICATION

In this paper, Fisher linear discriminant analysis (FLDA) is used to classify the EEGP-300 signals which are extracted from brain activities. In this case, at first the preprocessing algorithms such as filtering and referencing are applied to the raw EEG signal. Then, in order to create a model out of the signal, a linear predictive coding model with 6 order is used. So that the signal is reconstructed by extracting linear predictive coding (LPC) model parameters of each single trial, and then every signal trial is passed through the Hamming window by length 9. At last Fisher Linear Discriminant Analysis is used for classifying. In this paper, classification accuracy, the maximum bit rate and the convergence time to achieve stability in maximum accuracy of classification are computed to compare performance of the proposed method, Fisher Linear Discriminant Analysis with Linear Predictive Coding Model and Hamming Window (LPC+HAMMING+FLDA), to FLDA and LPC+FLDA. The implementation results show that the efficiency of the proposed method in terms of classification accuracy and convergence time to achieve stability in maximum accuracy is better than the other two mentioned algorithms. As example, at the proposed algorithm with 8 electrode configuration the S2 converges to the maximum accuracy after eleventh Block while this happens for two other algorithms after fourteenth Block and the total classification accuracy for this person at proposed algorithm is improved as 2.2% and 4% than respectively LPC+FLDA and FLDA algorithms.