Learning Behavior Analysis Based on Instant Message and Online Learning Platform

More and more teachers tend to use online learning platforms to assist teaching. Meanwhile, the instant messaging has generated increasing awareness of its educational value, due to its flexibility, convenience and widespread use. Many researches are to model and analyze the learning behavior based on data collected from online learning platforms or instant messaging (IM) platforms separately. But relatively little is known about how the students behave when combining these two kinds of platforms with traditional classroom learning. In this study, we investigate and analyze learning behaviors of first-year university students in a project-based course by exploring data collected from an online learning platform and an IM platform which are used in the course. Firstly, an interactive social network is constructed based on the IM messages, from which students’ engagement in the course and their preferred collaborative learning styles can be observed. Then learning behavior features extracted from both platforms are analyzed to get insights of the learning behaviors of students when combining these two kinds of platforms in the course. Our findings reveal that instant messaging platform can help students adapt to new fields, and their learning patterns on the IM platform are different, which ultimately leads to different learning performance. Finally, a learning performance prediction model called Student Predict Based On BI-LSTM (SPBI-LSTM) is proposed, which fuses student behavior sequences from two platforms to predict students who under-performing. We experimentally verify the improvement in accuracy when integrating instant messaging data.