Key points are not available for this paper at this time.
In the contemporary educational landscape, data-driven decision-making has become pivotal for enhancing student success. This article explores an intelligent analytic framework leveraging Multiple Linear Regression (MLR) and Random Forest (RF) algorithms to predict student performance, providing a comparative analysis of their predictive capabilities. MLR, a statistical technique, models the relationship between students' grades and various factors such as attendance and socio-economic background, offering transparency and interpretability of the impact of each predictor. RF, an ensemble learning method, excels in handling large datasets and capturing non-linear interactions among variables, offering higher accuracy in prediction. The study was conducted using 664 datasets from eight departments of Federal Polytechnic Ukana, following rigorous data preprocessing and normalization. The performance of both models was evaluated based on metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), R-squared Score (R²), and Explained Variance Score (EVS). The results revealed that RF outperformed MLR significantly, with lower error rates and higher predictive accuracy. Scatter plots and bar charts further illustrated the robust performance of RF over MLR. This research underscores the potential of integrating advanced machine learning techniques in educational settings to provide deeper insights into student performance, enabling timely and targeted interventions. The findings advocate for the adoption of RF for more accurate predictions and improved educational outcomes. Future research should explore hybrid models and expand the dataset to validate the applicability of these findings across diverse educational contexts.
Johnson et al. (Fri,) studied this question.