AIoT-Based Drainage Blockage Early Warning System

Student stress and burnout are major educational problems for college students, who often work under heavy demands such as examinations, large workloads, and tight deadlines. As a result, students’ well-being is seriously threatened. Usually, the problem is spotted only after a student’s performance declines, with performance being the only indicator. This paper reviews eight studies (2021–2026) that examine the potential of Artificial Intelligence and Machine Learning in detecting student burnout and stress. The studies cover explainable machine learning, the demands–resources model of burnout, questionnaire and behavioural data used as algorithm inputs, data obtained through wearable sensors, and passive smartphone sensing. Their approaches, results, and limitations are compared. Overall, the findings show that machine learning can be used to classify or forecast stress states. However, many studies rely on grade or self-reported data, use very small datasets, require specialised equipment, or raise privacy concerns. If these issues are addressed first, such systems could become a practical part of higher education and a valuable aid to student mental health support.