Predicting the Performance of Mathematics’ Students through Data Mining Techniques for Enhanced Education Systems

In this study, we use data mining techniques to predict the performance of students in mathematics. The failure rate in secondary education within the discipline of mathematics is significant. Therefore, predicting the performance of students is relevant to support the decision plan on interventions. A dataset with 395 records collected from two secondary schools in Portugal was used in this work. The authors tested eight machine learning algorithms such as Stochastic Gradient Descent, AdaBoost, Support Vector Machine, Naïve Bayes (NB), K-Nearest Neighbors algorithm, Decision Tree, and Logistic Regression. Consequently, this work supports future research activities to use NB to predict students’ performance. Moreover, this work has ranked the impact of the variables which most disturb students from achieving success. The results recommend the NB algorithm using the top five important features which report 75.22%, 74.41%, 75.22%, and 73,29% of accuracy, precision, recall, and F1-Score, respectively. By predicting students’ performance is possible to plan interventions on time and, consequently, increase the final discipline score.