Predicting Graduation Timeliness: A Multidisciplinary Study on Logistic Mobility, Psychosomatic Fatigue, and AI-Assistance Dependency among Information Systems Students

Timely graduation is now being impacted by non-academic stressors and technology use. This study goes beyond traditional models that rely solely on administrative data by incorporating physical and psychological dimensions, specifically Logistics Mobility (distance and duration of commute), Psychosomatic Fatigue, and AI Assistant Dependence. The data of 292 final year Information Systems students was balanced using Random Over Sampler. Models were compared between Random Forest and Support Vector Machine. Psychosomatic tiredness and AI assistant Dependence was the best predictive factor of delayed graduation, particularly for students commuting more than 25 km, who show more fatigue and higher AI reliance. In terms of algorithmic performance, Random Forest achieved an exceptional accuracy of 100%, while SVM followed closely with 98.28%. The results show the importance of physical exhaustion and digital dependency to identify students at risk of delay. The study finds that a student’s environment, in addition to cognitive ability, is a factor in whether or not they graduate on time. High reliance on AI indicates burnout, underscoring the importance of hybrid mentoring and AI literacy programs, particularly for commuting students.