Comparative Analysis of Machine Learning Models to Predict Depression, Anxiety and Stress

ABSTRACT

Mental Health Issues are a hidden pandemic which will emerge in the upcoming years. As the world witnessed COVID-19 pandemic and went into lockdown, the cases of Depression, Anxiety and Stress skyrocketed than ever before. This has given rise to the need for exploring the interdisciplinary field of Artficial Intelligence and Psychometry. In this paper, we propose compare various machine learning and ensemble learning methods, on the survey dataset comprising of the DASS-42 Psychometric Test Results and Demographic information. Random Forest, Decision Tree, Support Vector Machine (SVM), AdaBoost, CatBoost, and Extreme Gradient Boosting (XGBoost) are used to classify the level of Depression, Anxiety and Stress into normal, mild, moderate, severe and extremely severe categories. In our experiments on the dataset, Support Vector Machine outperformed and reached a final F1-measure of 94%, 95% and 91% in the prediction of Depression, Anxiety and Stress, respectively. © 2022 IEEE.