A deep learning framework for predicting the student's performance in the virtual learning environment

Abstract

Nowadays predicting the student's performance in the virtual learning environment is considered a critical point as it includes some of the student learning activities such as the course registration, tasks submissions, exams, as well as all the virtual interactions that happen so all of these are considered as a fertile field for research. In addition, Deep learning which is under the umbrella of artificial intelligence played an important role in the prediction's domain. Consequently, the study focused to discuss the role of artificial intelligence in the e-learning system in general and specifically the role of deep learning in predicting the student's performance, and it found that most of the studies focused only on the dropout prediction and neglect the other performance features as well as they didn't focus on improving the quality of the dataset. Consequently, the study proposed a deep learning framework to predict the student's academic performance in the virtual learning environment taking into consideration the quality of the dataset in the preprocessing layer, based on the deep neural networks the proposed model achieved a high accuracy of about 91.29% and low loss value about 0.18 compared to the other studies which utilized the same dataset.