Discovering genres of online discussion threads via text mining

摘要: As course management systems (CMS) gain popularity in facilitating teaching. A forum is a key component to facilitate the interactions among students and teachers. Content analysis most popular way study discussion forum. But content human labor intensity process; for example, coding process relies heavily on manual interpretation; it time energy consuming. In an asynchronous virtual learning environment, instructor needs keep monitoring from order maintain quality of However, consuming difficult instructors fulfill this need especially K12 This research proposes genre classification system, called GCS, automatic process. We treat as document task via modern data mining techniques. The posting can be perceived announcement, question, clarification, interpretation, conflict, assertion, etc. examines coherence between GCS experts' judgment terms recall precision, discusses how we adjust parameters improve coherence. Based empirical results, adopts cascade model achieve evaluation classified genres repository postings online earth science senior high school shows that effectively process, proposed deal with imbalanced distribution nature postings. These results imply based perform system.