Abstract—With the development of MOOCs, millions of
students enrolled into online courses. The discussion forums in
MOOCs provide a virtual community for students to interact
with each other. The communication in different topics indicates
the engagement of students in the courses and social-learning
process during the interactions. In this respect, this paper
explores the use of Naive Bayesian classification approach for
predicting the participation of the forum and the use of
Bayesian-based social-learning approach for modelling the
opinion formation process during the discussion and indicating
the influence of instructors in the discussion forum. Results on
data from 1 Coursera course demonstrate that the poster’s
retention can be well predicted by Naive Bayes classifier based
on the combination of different features of the forum postings;
additionally, we find that the superposters may not be the
participants who will continue posting in the last several weeks.
In terms of social-learning, our analysis indicates participants
will aggregate information by repeated interactions and the
instructors’ post can improve the convergence of learning
process to the true belief. These results confirm the influence of
the instructors’ intervention further.
Index Terms—Discussion forum, MOOCs, participation
prediction, opinion formation.
Tieying Zhu, Wei Wang, and Wei Zhao are with School of Computer
Science, Northeast Normal University, China (e-mail: zhuty@nenu.edu.cn,
wangw@nenu.edu.cn, zhaow@nenu.edu.cn).
Riming Zhang is with School of Software Engineering, Northeast Normal
University, China (e-mail: zhangrm281@nenu.edu.cn).
Cite: Tieying Zhu, Wei Wang, Wei Zhao, and Riming Zhang, "Participation Prediction and Opinion Formation in MOOC Discussion Forum," International Journal of Information and Education Technology vol. 7, no. 6, pp. 417-423, 2017.