Smart Jump: Automated Navigation Suggestion for Videos in MOOCs.

Statistics show that, on average, each user of Massive Open Online Courses (MOOCs) uses 'jump-back' to navigate a course video for 2.6 times. By taking a closer look at the navigation data, we found that more than half of the jump-backs are due to the 'bad' positions of the previous jump-backs. In this work, employing one of the largest Chinese MOOCs, XuetangX.com, as the source for our research, we study the extent to which we can develop a methodology to understand the user intention and help the user alleviate this problem by suggesting the best position for a jumpback. We demonstrate that it is possible to accurately predict 90% of users' jump-back intentions in the real online system. Moreover, our study reveals several interesting patterns, e.g., students in nonscience courses tend to jump back from the first half of the course video, and students in science courses tend to replay for longer time.