Personalized Stress Mitigation Through EEG Based Stress Classification and Music Recommendation

Stress has become a common and pervasive aspect of daily life affecting individuals across the world. Prolonged exposure to stress increases the risk of various physical and mental disorders. In this implementation paper, we propose an approach for stress alleviation by using electroencephalogram (EEG) based music recommendation. The primary objective of this study is to develop a web application which can accurately detect the stress levels and suggest relevant music to the individuals based on their stress levels.This study utilized EEG Brainwave dataset and employed machine learning algorithms, such as K-Means Clustering followed by Support Vector Machine (SVM) in order to classify stress levels into three categories such as low, medium and high. It became evident from our results that the SVM classifier is effective in detecting stress levels with accuracy of 95%. The results are then used to recommend relevant music to the person for managing the stress using music therapy. Overall, this implementation paper offers a useful solution for coping with stress by leveraging music’s therapeutic effects.