Machine Learning-Based Prediction of Students' Academic Performance in Higher Education
TL;DR
Seven machine learning methods were applied to classify and predict student academic performance: Multinomial Naïve Bayes, Support Vector Machines, Decision Trees, Logistic Regression, Random Forest, Gradient Boosting and k-Nearest Neighbors.
Abstract
In higher education, student academic performance is considered a key educational outcome. Consequently, improving academic performance is a pressing objective for many higher education institutions. Predicting and classifying student performance can provide valuable insights and facilitate the development of appropriate strategies to enhance academic results. In recent years, data mining methods have been widely applied and have demonstrated significant success in classification and prediction tasks. This article presents and analyzes the EASP student performance dataset. Seven machine learning methods were applied to classify and predict student academic performance: Multinomial Naïve Bayes, Support Vector Machines, Decision Trees, Logistic Regression, Random Forest, Gradient Boosting, and k-Nearest Neighbors. The performance of these methods was evaluated using precision, recall, the area under the ROC curve (AUC), and the F1-score. Experimental results demonstrated that Logistic Regression achieved the best predictive performance on the EASP test set, with a precision of 0.75.