A Personalized Virtual Learning Environment Using Multiple Modeling Techniques

Student learning optimization is one of the main goals of education. A conventional e-learning system fails to accomplish its true purpose due to the lack or absence of personalization features. This paper presents an end-to-end approach for supporting students’ diverse needs by classifying their learning styles in a virtual learning environment (VLE) and embedding the discovered knowledge in an adaptive e-learning system prototype. Furthermore, we validated different models’ accuracies and comparative consistencies to manual methods using 704,592 interactions log data of 898 learners. Quantitative results show that the Support Vector Machine (SVM) achieves cross-validated accuracies of 88%, 86%, and 87% (processing, perception & input) of the Felder-Silverman Learning Style Model (FSLSM) and the Decision Tree (DT) for the understanding dimension with 86% accuracy.