Toward Improved Classification of Perceived Stress Using Time Domain Features

Abstract

Perceived stress is the predominant mental health concern in this age of development and progress. This research study aims to classify perceived stress using non-invasive electroencephalography (EEG) signals. The dataset employed in this research comprises EEG information from twenty-eight participants in a closed-eye state, utilizing commercially available Muse EEG headbands. We have preprocessed EEG data and performed analysis on EEG data spanning 210 seconds. Two segmentation techniques were employed: non-overlap and overlap. The information-gain-based feature selection method was applied before classification to extract distinct features to enhance feature relevance and reduce dimensionality. To categorize the EEG data into stressed and non-stressed groups, the Perceived Stress Scale (PSS) questionnaire was utilized. Employing a Random Forest classifier alongside an overlap segmentation technique, our proposed method attained a peak classification accuracy of 93.8%. This surpasses existing stress classification schemes found in the literature. INDEX TERMS classification, electroencephalography, feature extraction, Perceived stress, time domain features.