A data mining driven approach to real-time classroom feedback through multimodal sensing

The communication of complex concepts is a challenge for teachers, not only in primary and secondary schools but in university classrooms. In order to adapt and maximize the impact of the lectures on the class, it is necessary that teachers receive frequent and reliable feedback from their students. Currently, such feedback is typically gathered through written quizzes or surveys, which cannot accurately measure levels of student engagement in real time over the course of a full lecture. In this paper, we present a novel data mining driven approach for gathering classroom feedback by digitally analyzing student body language. We use a multimodal sensor to precisely record students’ body language in real time, and use classification rule mining to compare the body language data with feedback obtained with written surveys. From the methodology proposed, insights will be gained into the feasibility of using digital body language analysis to accurately assess levels of student engagement.