The phenomenon of "student burnout" has become an important consideration in higher education due to its negative consequences for mental health, academic achievement, an‎d the long-term development of students. Research shows that student burnout often occurs early in undergraduate education an‎d, therefore, early intervention is necessary to assist students impacted by burnout. This study attempts to address this gap by proposing a predictive model that integrates machine learning algorithms with educational data mining (EDM) an‎d can detect students at risk for burnout before the onset of severe symptoms. This work employs Support Vector Machines (SVM) an‎d Ran‎dom Forest classification algorithms with clustering an‎d Association Rule Mining (ARM) to analyze data obtained from academic records, general learning management systems (LMS) activity, an‎d psychometric self-report questionnaires using the Maslach Burnout Inventory–Student Survey (MBI-SS). Additionally, this study introduces two new predictive constructs, engagement score an‎d studentship score, to account for cognitive, behavioral, an‎d social dimensions of student functioning. The addition of these features significantly contributed to the predictive accuracy an‎d interpretability of the model to better inform targeted, an‎d personalized early intervention practices. Additionally, the findings of this study demonstrate that patterns of risk for student burnout may vary significantly by discipline showing that contextual an‎d faculty variables are important for the accuracy of predictive modeling. This study thus offers a contribution to the development of proactive, mental health supports in higher education an‎d the possibilities of finding the most defensible, practical, an‎d effective means of incorporating machine learning into the psychometric an‎d behavioral aptitude analytics that could provide early detection of burnout. The proposed model offers institutions a scalable, an‎d data-driven way to coordinate specific acts that benefit student well-being an‎d retention.

10/28/2025 12:00:00 AM

88199

1404/08/13

Early Detection of Student Burnout in Higher Education Using Data Mining

Student Burnout , Early Detection, , Maslach Burnout Inventory (MBI-SS) , Machine Learning , Educational Data Mining, Student Engagement , Predictive Analytics