Abstract Online learning communities provide open workspaces allowing learners to share information, exchange ideas, address problems and discuss on specific themes. But with the continuously increasing artifacts in online communities, learners feel it difficult to quickly and easily gain an insight into a certain theme. To facilitate and support learners have a better understanding of the communication focus, this paper presents an approach to discover the hot topics and patterns of topics evolutions in online learning communities. Firstly, hot terms are extracted based on three features: the frequency of the terms used in the document collection; the location of the terms within a document; the breadth of terms distribution in the document collection. Then a term association network is constructed by computing the terms co-occurrence and distance between them. Finally, an algorithm is proposed to select the kernel term and its associated terms as term clusters to represent the hot topics with multi-facets expression. Two case studies on real datasets are conducted to demonstrate the effectiveness and usefulness of term cluster in helping users better understand hot topics in online learning communities. Potential applications in learning scenarios are also discussed.
展开▼
