Clustering Analysis of Learners’ Watching Sequences on MOOC Videos

In MOOC community, it is a challenge to analyze online learners’ video watching behaviors based on the MOOC videos’ log data. In this paper, we propose a research model on video watching behaviors of MOOC learners, by encoding the behavior sequences retrieved from the videos’ log data, evaluating the similarity between different online learners, clustering and visualizing the similarity results. To verify our findings, we present a case study on four courses from a popular Chinese MOOC platform. In particular, we divide the learners into 5 different clusters based on the learners’ behavior patterns. The clustering results show that the video watching coverage and completion rate are key attributes that affect the clustering of learners. Besides, the ordering and distribution of the learners’ different activities are also important factors that decide the learning patterns in each cluster. The distribution of learners is also affected by different types of courses in each cluster. Based on the exercise log data in the same MOOC platform, we further calculate average assessment scores in each cluster, which reveal obvious differences across different clusters.