Student Health Risk Evaluator (SHRE): A Predictive Machine Learning Framework for Assessing Mental Health Risks in Students

Mental health problems in students pursuing higher education are on the rise and affect learning participation, performance, and well-being. The traditional methods such as self-reporting surveys or face-to-face interviews are restricted in their ability to provide timely information to educators and policymakers. This paper introduces the Student Health Risk Evaluator (SHRE), an artificial intelligence (AI)-based predictive analytics model that can assist the counsellors and educators to evaluate mental health risk of the students pre-emptively and take timely preventive measures. Data was gathered using questionnaire covering major indicators to assess mental health like personal well-being, emotional and psychological health, social behavior, etc. Lasso regularization was utilized using two different alpha values (0.05 and 0.1) to identify the most influential predictors affecting mental health of the students. The selected features were then used to train and evaluate machine learning algorithms such as Logistic Regression, Support Vector Machine, K-Nearest Neighbors, and Random Forest, which were applied on the dataset to predict the mental health of the students. Logistic Regression achieved the maximum accuracy of 94.12% and F1-score of 93.98 at an alpha value of 0.05 amongst all others. The findings demonstrate the potential of AI-driven models to support early identification of mental health risks and offer a data-driven foundation for timely interventions. While the current study focuses on questionnaire-based prediction, future work may explore the integration of additional data sources and broader institutional deployment to enhance its practical applicability.