A prediction model based machine learning algorithms with feature selection approaches over imbalanced dataset

The educational sector faced many types of research in predicting student performance based on supervised and unsupervised machine learning algorithms. Most students' data are imbalanced, where the final classes not equally represented. Besides size dataset, this problem affects model's prediction accuracy. In paper, Synthetic Minority Oversampling Technique (SMOTE) filter is applied to dataset find its effect Four feature selection approaches most correlated attributes that affect performance. SMOTE examined before after applying measure accuracy with Three supervised/unsupervised algorithms predict findings show (LMT, Simple Logistic, Random Forest) got high without selection. accuracies (Canopy, EM, Farthest First) enhanced filter.