Predicting Student Attrition in MOOCs using Sentiment Analysis and Neural Networks.

Abstract

While there is increase in popularity of massive open online courses in recent years, high rates of drop-out in these courses makes pre- dicting student attrition an important problem to solve. In this paper, we propose an algorithm based on articial neural network for predict- ing student attrition in MOOCs using sentiment analysis and show the signicance of student sentiments in this task. To the best of our knowl- edge, use of user sentiments and neural networks for this task is novel and our algorithm beats the state-of-the-art algorithm on this task in terms of Cohen's kappa.