Improving classification accuracy of EEG based brain computer interface signals

Feature extraction is a very crucial step at modern electroencephalogram (EEG) based brain computer interface system. Various feature extraction techniques have been proposed in order to represent EEG signals. With this study, it was shown that the classification accuracy increased by extracting features from different time segment of EEG signals. The proposed method improved the average classification accuracy to 69.08% which was 65.35% at the previous study.