Abstract

Depression is a significant and growing concern within academic environments, affecting both students and staff due to factors such as academic pressure, financial stress, and lifestyle challenges. This study explores the use of machine learning, specifically a Random Forest classifier, to predict depression risk among students using behavioral, psychological, and demographic data. A dataset of 27,788 student records was analyzed after thorough preprocessing and exploratory data analysis. The model achieved strong performance, with an accuracy of 83.52% and an AUC of 0.91, indicating reliable classification of depression status. Local Interpretable Model-agnostic Explanations (LIME) were employed to enhance interpretability, revealing key predictive features such as suicidal ideation, academic pressure, sleep duration, and dietary habits. These interpretable insights align with existing psychological research and provide actionable information for mental health professionals. The findings highlight the value of explainable AI in educational settings, offering a scalable and transparent approach to early depression detection and intervention. Future work should focus on longitudinal data integration, multimodal inputs, and real-world implementation to strengthen the model’s utility and impact.