Machine and Deep Learning Models for Stress Detection Using Multimodal Physiological Data

Stress detection using physiological data has gained significant attention due to its applications in healthcare and well-being. This project explores machine and deep learning models to classify stress levels based on multimodal physiological signals. Four models—Multilayer Perceptron (MLP), Random Forest, Decision Tree, and Logistic Regression—are implemented and evaluated for their performance in stress prediction. The study leverages feature extraction and preprocessing techniques to enhance model accuracy, comparing their effectiveness in distinguishing between stress and non-stress states. Experimental results demonstrate the comparative strengths of each model, with tree-based algorithms like Random Forest and Decision Tree showing robust performance, while MLP provides deeper learning capabilities for complex patterns. The findings highlight the potential of machine and deep learning in automated stress detection systems, contributing to real-time mental health monitoring. This research underscores the importance of model selection and optimization in improving classification accuracy for physiological stress analysis.

Keywords: Stress detection, Machine Learning, Deep Learning, Multimodal physiological data, MLP, Random Forest, Decision Tree, Logistic Regression, Classification, Mental health monitoring.