Massive Open Online Course (MOOC) platforms such Coursera and edX have started to integrate education technology and learning analytics applications to let instructors manage the class effectively and get feedback on the class performance. A key aspect of class management is for instructor team (instructor(s) and his teaching assistants (TAs)) to talk back to the students who express their thoughts and post their queries via discussion forums. However due to the class size of a typical MOOC instructors are unable to participate in all student discussions or answer all of their queries. We are building content based recommender systems to rank the discussions threads by their priority and importance to help to the MOOC instructor.
The output from the models need to be presented to the MOOC instructor in a dashboard. The dashboard should have a generic backend API to consume discussion forum posts from different MOOC and e-learning platforms such as Coursera, edX, Piaza and NUS IVLE. We already have a backend to can crawl
and consume Coursera?s discussion forums. The API should then feed the posts to our machine learning models [1,2]. The output from the models should be used to display the forum threads in a new order different from that originally displayed on the MOOC platform.
Proceedings of the Thirty-First AAAI conference on Artificial Intelligence (AAAI-17), 2017.
Proceedings of the International Conference on Information Systems (2015), 2015.
- Divish Dayal (Undergraduate Intern – Alumni)
- [insert_php] echo get_avatar( $id_or_email=’email@example.com’, $size=30 ); [/insert_php] Zhang Yichi (Undergraduate FYP Student – Alumni)
- [insert_php] echo get_avatar( $id_or_email=’firstname.lastname@example.org’, $size=30 ); [/insert_php] Min-Yen Kan (Professor and Advisor)
- [insert_php] echo get_avatar( $id_or_email=’email@example.com’, $size=30 ); [/insert_php] Muthu Kumar Chandrasekaran (Project Lead)
- You can process your Coursera’s SQL data dumps using our free, open source library lib4MOOCdata. This also serves as a codebase to replicate our published findings.