DETECTION OF MENTAL STRESS USING BIOSIGNALS THROUGH MACHINE LEARNING- A BRIEF SURVEY

Abstract

This paper explores the classification of various existing biosignals used for stress detection and evaluates their effectiveness. Out of all the biosignals reviewed, ECG was chosen as the best  biosignals . Further review was conducted to determine the most suitable machine learning model  for the chosen ECG signal. Results have shown that KNN gives the best accuracy of 96.41%  followed by SVM with a considerably high accuracy of 90.10%. The best ML models were then  evaluated on other biosignals for comparison of effectiveness on different data signals. The  findings of this study offer useful insights into the selection of optimal biosignals and machine  learning algorithms for detecting stress, which can contribute to development of personalized  stress management technologies and improving mental health.