MOOC Video Personalized Classification Based on Cluster Analysis and Process Mining

In the teaching based on MOOC (Massive Open Online Courses) and flipped classroom, a teacher needs to understand the difficulty and importance of MOOC videos in real time for students at different knowledge levels. In this way, a teacher can be more focused on the different difficulties and key points contained in the videos for students in a flipped classroom. Thus, the personalized teaching can be implemented. We propose an approach of MOOC video personalized classification based on cluster analysis and process mining to help a teacher understand the difficulty and importance of MOOC videos for students at different knowledge levels. Specifically, students are first clustered based on their knowledge levels through question answering data. Then, we propose the process model of a group of students which reflects the overall video watching behavior of these students. Next, we propose to use the process mining technique to mine the process model of each student cluster by the video watching data of the involved students. Finally, we propose an approach to measure the difficulty and importance of a video based on a process model. With this approach, MOOC videos can be classified for students at different knowledge levels according to difficulty and importance. Therefore, a teacher can carry out a flipped classroom more efficiently. Experiments on a real data set show that the difficulty and importance of videos obtained by the proposed approach can reflect students’ subjective evaluation of the videos.