Analysis of EEG signal to classify mental state using stacking ensemble classifier.

Classification of a person's cognitive state is important in many applications. The aim of this work is to develop a reliable classifier for predicting the cognitive state of the brain while performing various tasks. This dataset was recorded using a Muse headset with 10-20 EEG electrode placements and four different electrodes. The cognitive state of ten students was recorded while performing various tasks, such as mental arithmetic, reading technical articles, listening to technical podcasts, surfing the Internet, and resting with eyes open or closed. The EEG dataset is preprocessed and feature sets of five signals alpha, beta, gamma, delta and theta are selected for prediction. This dataset is trained by any machine learning algorithm such as decision trees, naive Bayes, and support vector machines. All machine learning algorithms are then embedded into a Stacking Ensemble Classifier with 10-fold cross-validation to predict the cognitive state of brain signals and classify them as high and low. Compared with existing algorithms, the method achieves an overall accuracy rate of 96%. [ABSTRACT FROM AUTHOR]