Student Performance Prediction Using Machine Learning: A Comprehensive Analysis

The anticipation of student performance stands as a pivotal element in educational systems, with organizations aspiring to enrich the learning experience and elevate student outcomes. Its prominence in the education domain arises from its capacity to refine educational results and furnish invaluable insights for educators, administrators, and policymakers alike. In this paper, we use the Student Performance Dataset (SPD) to evaluate the effectiveness of different machine learning methods across diverse application scenarios. More precisely, we explore the performance of eighteen machine learning models that include classification models, namely Artificial Neural Networks (ANNs), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Naive Bayes (NB), Decision Trees (DT), Random Forests (RF), Boosting and Bagging models. These methods are systematically applied to analyze binary prediction tasks within the context of student performance. Various machine learning algorithms, are employed to analyze and predict student performance metrics, such as grades, exam scores, and graduation outcomes. Evaluation of predictive models is a critical aspect, and the paper examines various performance metrics such as Accuracy, Precision, Recall, F1-measure, and the Area Under the Receiver Operating Characteristic curve (AUC). The experimental results demonstrate that the Categorical Boosting model (CatBoost) outperformed the rest of the models used in the study as the best-performing model in general, as it consistently achieved high scores in accuracy, recall, F1 score, and AUC. The results also showed that the results of the Decision Tree (DT) model were lower than ensemble methods, indicating potential limitations in handling complex relationships. In addition, the performance of Bagging techniques generally improved performance compared to their base models, demonstrating the effectiveness of aggregating multiple models and Boosting Techniques models consistently performed well, indicating the power of sequential learning and model combination. Keywords: Machine learning Models, Supervised Learning, Classification Algorithms, Student performance prediction