For access to this article, please select a purchase option:
IET members benefit from discounts to all IET publications and free access to E&T Magazine. If you are an IET member, log in to your account and the discounts will automatically be applied.
Your recommendation has been sent to your librarian.
Nowadays, because of the widespread of the online learning platform, automatic engagement detection arise a lot of researchers' interest. With the help of engagement detector, teachers or platforms can arrange the learning materials more personally for each learner. Currently, most work only focus on improving the accuracy of the engagement detector. While this work think only detect the engagement level is not enough, we also need to consider how to actual use it in onlinelearning platform. To this end, we first introduce the balanced random forest model to solve the classification task, then a method based on forest model is designed to find its local explanation (ie. why the student is disengaged). We think the local explanation can provide a reference for lecturers or platforms to improve the learners' learning experience. The system is tested through DAiSEE dataset. The final F1-score reaches to 0.809 for detector. We then compare our local explanation with LIME[1], the top 5 result are almost the same. However, our method is fast enough to be used in an realtime system. The result shows that our method does work.
Inspec keywords: educational administrative data processing; random forests; computer aided instruction; behavioural sciences computing
Subjects: Combinatorial mathematics; Computer-aided instruction; Social and behavioural sciences computing; Educational administration