Text Enhanced Ensemble for Human Stress Prediction

This research explores machine learning techniques for predicting human stress levels. Two data processing strategies, Bag of Words and TF-IDF, were employed prior to using the XGBClassifier for prediction. The results revealed that the Bag of Words strategy outperformed TF-IDF. Furthermore, the XGBClassifier demonstrated strong predictive power, and the Logistic Regression classifier also displayed promising performance. To further enhance accuracy, we implemented an ensemble model known as the Voting Classifier, which combined the strengths of multiple classifiers. The ensemble approach achieved the highest accuracy, highlighting the potential of combining classifiers for improved stress prediction. These findings offer valuable insights for addressing stress-related challenges in today’s fast-paced society.