Effects of PCA-Enabled Machine Learning Classification of Stress and Resting State EEGs

This paper aims to investigate the effects of Principle Component Analysis (PCA) enabled approach for the classification of stress and resting state of Electroenphalogram (EEG) when performing Mental Arithmetic Tasks (MAT) through the usage of signal processing and machine learning classifiers. For this analysis, we investigate our approach on a group of 25 subjects from the SAM-40 dataset and subsequently using the average power computed as a feature for our machine learning classifier. The results have shown the benefits of having PCA-enabled features where an increase of more than 10% is achieved, paving way for the potential future usage as a computer-aided diagnosis tool to assist in the detection of stress.