Early Prediction of At-Risk Students in a Virtual Learning Environment Using Deep Learning Techniques

Abstract

AbstractWith the advancement of the internet and communication technologies, online learning has gained acceleration. The largely-scaled open online courses run on specific virtual platforms, where learners can engage themselves in their own space and pace. The Virtual Learning Environments (VLE) have shown rapid development in recent years, allowing learners to access high-quality digital materials. This paper aims at exploring students' affinity towards early withdrawal from online courses. The work expands by finding learner-centric factors contributing to students' early prediction at-risk of withdrawal and developing a prediction model. The current work uses the free Open University Learning Analytics Dataset (OULAD). Here, the early identification of students at risk of withdrawal is predicted based on a Deep Learning Approach using CNN Algorithm. Time-series analysis is done using data from consecutive years. The work's significant contribution is a set of influential parameters predicting at-risk students at an early learning stage. The prediction accuracy falls in the range of 83% to 93%.KeywordsLearning analyticsAt-risk student predictionStudent dropoutDeep learning