HUMAN STRESS DETECTION USING MACHINE LEARNING

Stress is frequently characterized as a mental or emotional state provoked by challenging or unavoidable circumstances, referred to as stressors. Comprehending stress levels in individuals is essential for averting adverse consequences in life. Sleep disorders are linked to various medical, emotional, and social issues. This study seeks to investigate the capability of machine learning algorithms in identifying human stress through sleep-related behaviors. The dataset includes diverse sleep patterns and stress levels. Following data preprocessing, six machine learning algorithms—Multilayer Perceptron (MLP), Random Forest, Support Vector Machine (SVM), Decision Trees, Naïve Bayes, and Logistic Regression—were utilized for classification to attain accurate results. The results of the experiment indicate that the Naïve Bayes method emerges with the lowest mean absolute error (MAE) and root mean squared error rates (RMSE). This technique can classify data with an impressive accuracy of 91. 27%, illustrating high precision, recall, and F-measure values. These results are invaluable for evaluating human stress levels and addressing associated issues in a timely manner. KEYWORDS: Stress detection, Machine learning, multilayer perceptron (MLP), Random Forest, Decision tree, Gradient boosting, Naïve Bayes.