On Modeling LMS Users’ Quality of Interaction Using Temporal Convolutional Neural Networks

Abstract

Learning Management Systems (LMSs) have been widely employed following the Covid-19 pandemic. The user modeling of LMS including educators and learners is a point of interest for Higher Education Institutions (HEI), stakeholders and system users. In this work user’s engagement with LMS is modeled using the Quality of Interaction (QoI) indicator under a combined approach of blended and collaborative learning. The present research extends the previous work of ‘Fuzzy QoI’ and ‘DeepLMS’ to develop a generalized model that substitutes the fuzzy logic system with a deep learning model. In this line, Temporal Convolutional Neural Networks (T-CNN) were used to predict QoI, achieving MAE (0.027), RMSE (0.066) and R2 (0.698). The feedback received from the T-CNN model provides insights to educators and stakeholders in order to enhance the pedagogical experience.