Scaffold Learners? Qualitative Browsing of Major Topics in Learning Forum

After the system database contains an accumulation of discussion messages with date tagged from different classes within a course, an instructor may want to have an overview of learners’ discussions, such as the top ten hot topics, without reading every message in the learning forum. Hence, it is necessary to model the relationships between discussion messages during a specific time interval for showing the major discussion topics. Consequently, an instructor can use the information to refine his/her teaching strategies in the classroom or virtual classroom. This paper proposes a statistical model of word usage, which originates from information retrieval technology, to group learners’ discussion messages by concept similarity into time dependent topics. Consequently, the methodology can generate comprehensible topics for an instructor to sketch out what learners were discussing by concept-based message groups and when the important messages were posted by a temporal interface.