ML pipeline classifying stress vs. baseline state from wearable ECG, EDA, and skin temperature signals (WESAD dataset) using Random Forest, SVM, and KNN with subject-independent (LOSO) validation.

An end-to-end machine learning pipeline that classifies Baseline, Stress, and Amusement affective states from chest-worn physiological signals using the WESAD dataset. Features subject-wise validation to prevent data leakage and automated relative path tracking for portable deployment.