Predicting Students’ Performance using Machine Learning Classifiers

This paper presents students’ performance prediction model based on Machine Learning (ML) classifiers. The used dataset provides three groups of features that may affect students’ performance. Features selection approaches have been applied to the dataset. Several well-known ML classifiers including Neural Network (ANN), Bayes Network (Bayes Net), and Decision Tree (DT). Along with the original approach, hyper approach is applied using boosting, bagging, and voting classifiers. The performance of these algorithms is examined and compared using a confusion matrix. Results show that behavioral features have a major impact on students’ performance and ML algorithms can be used effectively to build students’ performance prediction model.