Confusion detection using neural networks

Abstract

Educational data mining (EDM) using enhanced research methods are allowing researchers to effectively model a spectrum of paradigms affecting students learning, including various epistemic emotions like confusion. confusion plays a vital role in learning, and some amount of confusion is constructive in learning new knowledge. However, when confusion is left unattended for long, it may lead the student to lose interest or feel frustrated and eventually drop out of the course. In this paper, we investigate student’s performance to detect the level of confusion in the exercises they attempt online. We investigate the performance of feedforward neural network algorithm, MLP (Multi-Layer Perceptron), and report the results and comparison of various algorithms and how the same methodology can be extended to any Learning Management System (LMS) on various digital learning platforms, including MOOCs especially because they suffer from high drop-out rates. We also discuss how we plan to extend our research to include more features to make it appropriate for cross-domain implementation.