A COMPARATIVE EVALUATION OF SUPERVISED MACHINE LEARNING MODELS FOR PREDICTING ACADEMIC PERFORMANCE FROM STRUCTURED EDUCATIONAL DATA

Abstrak:

The increasing availability of structured educational data has enabled the application of supervised machine learning techniques to predict students’ academic performance. Accurate prediction models can support early intervention strategies, personalized learning, and institutional decision-making. This study presents a comparative evaluation of widely used supervised machine learning models—including Logistic Regression, Decision Trees, Random Forest, Support Vector Machines, k-Nearest Neighbors, and Artificial Neural Networks—for predicting academic outcomes using structured educational datasets. The comparison is conducted based on predictive accuracy, interpretability, computational complexity, and robustness, as reported in prior empirical studies. The findings indicate that ensemble-based methods and margin-based classifiers generally outperform simpler linear models in predictive accuracy, while simpler models retain advantages in interpretability and implementation. The study synthesizes existing evidence to guide researchers and educational practitioners in selecting appropriate machine learning models for academic performance prediction.