Decision Tree Model for Classifying University Students Eligible for UKT Waivers

Abstract

This paper develops a Decision Tree-based classification model to determine student eligibility for UKT (Single Tuition Fee) waivers using socio-economic factors such as parental income, household type, parental occupation, number of dependents, and vehicle ownership. The goal is to automate the identification of students qualifying for financial aid, improving efficiency and fairness in resource allocation. The model was trained on a dataset containing both categorical and numerical features, with the target variable being binary: "Eligible" (1) or "Not Eligible" (0). The model achieved an overall accuracy of 93.33%, with strong performance for the "Eligible" class, reflected by excellent precision, recall, and F1-score. However, the model performed poorly on the "Not Eligible" class, with low recall and F1-score, highlighting the issue of class imbalance. To address this, techniques like resampling and class weighting are recommended to improve classification of the minority class. Exploring alternative models like Random Forest or XGBoost could also provide more balanced results. This  underscores the importance of addressing class imbalance and using evaluation metrics beyond accuracy when developing classification models for imbalanced datasets.