Predicting Student Final Grades Using Random Forest Algorithms and Linear Regression

The increasing adoption of intelligent systems in higher education has encouraged the use of data-driven approaches to predict students’ academic performance. Accurate prediction models are essential to support early intervention and informed academic decision-making. This study aims to conduct a comparative analysis between Random Forest and Linear Regression algorithms in predicting students’ final academic scores. The dataset consists of assessment components, including quiz scores, assignment scores, and midterm examination (UTS) scores, which are used as predictor variables. The data were divided into training and testing sets with a ratio of 80:20. Model performance was evaluated using accuracy, error metrics, and feature importance analysis. The experimental results indicate that Random Forest outperforms Linear Regression in terms of predictive accuracy and robustness. Furthermore, both models consistently identify midterm examination scores as the most influential factor affecting students’ final performance. These findings demonstrate that ensemble-based learning methods are more suitable for academic performance prediction and can serve as a reliable foundation for intelligent academic support systems in higher education.