| Abstract | In online discussion platforms, participants gather to discuss the effective approaches to solve the common issues they face. To facilitate these discussions proceeding smoothly and reaching consensus efficiently, human facilitators are introduced into these discussions. However, human facilitator-related problems such as human bias and scalability arise with the increasing sophistication of these online discussions. As a result, it becomes critical to find approaches that support facilitation in these online discussions. Towards this end, we propose a novel case-based reasoning framework to support online discussion facilitation. In the proposed framework (CBR), each discussion thread is styled using an issue-based information system (IBIS), where complex problems are modeled as argumentation processes among several stack holders. Upon identifying a new discussion case, facilitation suggestions are generated by retrieving the most similar facilitated case from the discussion case base. For the sake of experimental evaluation, we build an online facilitation dataset that contains the facilitation information from annotated real-world discussion data. The experimental results show the validity of the generated facilitation suggestions. |
|---|
