A CASCADED BI-LEVEL FEATURE SELECTION TECHNIQUE FOR PREDICTING STUDENTS’ ACADEMIC PERFORMANCE

Abstract

ABSTRACT Educational Data Mining is an important task which is used to detect and explore useful patterns applicable to student learning behavior. Features in educational data are ambiguous which leads to noisy features and the curse of dimensionality problems. These problems can be solved via feature selection. There are existing models for features selection. These models were created using either a single-level embedded, wrapper-based or filter-based methods. However singlelevel filter-based methods ignore feature dependencies and also ignore the interaction with the classifier. The single-level embedded and wrapper based feature selection methods interact with the classifier, they can only select the optimal subset for a particular classifier. So the features selected by them may be worse for other classifiers. Hence this research proposes a robust a cascade bi-level feature selection technique for student performance prediction that will minimize the limitations of using a single-level technique, hence improve prediction performance. The proposed cascaded bi-level feature selection technique consists of the Relief technique at first-level and the Particle Swarm Optimization (PSO) at the second-level. The proposed technique was evaluated using the Eurostat student performance dataset. In comparison with the performance of the single-level feature selection technique the proposed technique achieved an accuracy of 94.94% for Mathematics dataset which was better than the 93.67% and 92.41% achieved by the single-level PSO and Relief selectors for Mathematics dataset for the binary classification task. The proposed technique also produced better results than previous works based on Eurostat dataset. These results shows that proposed bi-level cascade can effectively predict student performance.