Machine Learning and Differential Equation-Based Epidemic Model Framework for Predicting Student Burnout and Depression Due to Academic Workload During and Post-Pandemic.

Abstract

The rapid increase in academic workload during and after the COVID-19 pandemic has significantly affected the mental health of university students, leading to higher incidences of burnout and depression. Traditional statistical methods provide valuable insights into the relationships between academic and psychological variables but are often inadequate for capturing complex nonlinear interactions and forecasting the temporal progression of mental health conditions. This study proposes a novel Machine Learning and Differential Equation-Based Epidemic Model Framework for predicting student burnout and depression resulting from academic workload. Initially, statistical analyses, including descriptive statistics, correlation analysis, and regression techniques, are employed to identify significant academic, behavioral, and psychological factors influencing student mental health. Subsequently, an Artificial Neural Network (ANN) is developed to learn nonlinear relationships between these factors and to accurately predict individual burnout and depression levels. The ANN-predicted outcomes are then integrated into a compartmental epidemic-inspired differential equation model, where the transition parameters are dynamically estimated rather than assumed constant. The proposed mathematical model describes the temporal evolution of healthy, burnout, depressed, and recovered student populations under varying academic conditions. Numerical simulations are performed using synthetic and survey-based datasets to investigate the effects of workload intensity and intervention strategies on student mental health. The results indicate that combining ANN with differential equation modeling provides both high prediction accuracy and mathematical interpretability while enabling long-term forecasting of burnout and depression dynamics. The proposed framework offers an effective decision-support tool for educational institutions to implement early intervention strategies, optimize academic policies, and promote student psychological well-being.