Cognitive State Classification Using a single-channel Headset: An EEG Analysis Approach

A growing number of people are using electroencephalography (EEG) as a noninvasive way to record and monitor their brainwaves. An EEG is a technique that captures the electric signals generated by the brain using tiny metal discs connected to the scalp. The goal of this paper is to classify the cognitive state into two categories using information from the Neurosky headset, namely maths and relax label. In this paper, we have used logistic regression, linear SVM, non-linear SVC, random forests, XG boosting, MLP, gradient boosting, and neural network to make accurate predictions about the subject's cognitive state. The performance of the machine learning models were analyzed on the basis of various evaluation parameters like accuracy, precision, recall and F-1 score. The random forest algorithm had the highest accuracy rate of 89%, followed by the XG boosting algorithm, which had an accuracy rate of 87%. After tuning the hyper parameters, the XG boosting algorithm's accuracy increased by 3 and as a result, XG boosting model emerged as the best model, with a 90% prediction accuracy.