Forecasting the Academic Horizon: Machine Learning Models Unraveling the Complex Web of Student Well-being Determinants

In the contemporary academic landscape, the well-being of students is pivotal not only for their individual success but also for the broader educational ecosystem. This study meticulously delves into a rich dataset encompassing diverse student attributes, academic performance metrics, and economic indicators to discern patterns and predictors affecting student well-being. Leveraging a multi-faceted research methodology, we employed various machine learning models, ranging from logistic regression to advanced ensemble methods, aiming to forecast and comprehend the intricate determinants of student outcomes. The research design, underpinned by rigorous exploratory data analysis, revealed intriguing correlations between economic conditions, academic achievements, and students' well-being. The Gradient Boosting model, in particular, showed a significant improvement post hyperparameter tuning, with an accuracy reaching up to 77.63%. On the other hand, models like the Random Forest achieved a base accuracy of 77.29%. These insights highlight the potential of data-driven methodologies in understanding and predicting student well-being. As we stride into an era where data-driven decisions in education are paramount, our findings offer a robust foundation for future endeavors in this realm. Future directions of this study encompass refining prediction models with more granular data, exploring the psychological facets of student well-being, and devising actionable interventions based on the identified predictors.