Using Random Forest to Classify Financially Eligible Students for UKT

Abstract

This research investigates the use of a Random Forest-based classification model to automate the process of determining students' financial eligibility for the Uang Kuliah Tunggal (UKT) tuition assistance system in Indonesia. By leveraging socioeconomic data such as household income, family size, parental education level, and student performance, the model aims to enhance transparency, fairness, and efficiency in financial aid allocation. The dataset, comprising 1,000 student records with categorical and numerical features, was split into training (80%) and testing (20%) sets. The Random Forest model achieved a high overall accuracy of 90%, with exceptional performance for the Worthy class, attaining a recall of 100% and an F1-score of 0.94, ensuring no eligible students were overlooked. However, the model demonstrated lower recall (60%) for the Not worthy class, indicating room for improvement in addressing class imbalance. Key socioeconomic factors emerged as significant determinants, aligning with traditional UKT criteria. Future work should focus on enhancing model performance through data balancing techniques, feature enrichment, and exploring advanced machine learning algorithms. This research underscores the potential of data-driven approaches to improve the equity and efficiency of tuition assistance systems in higher education.