Tsuyoshi Murata, Tomoyuki Ikeya
Analyzing social interactions of CGM (Consumer Generated Media) is important for encouraging commnication among users. In the case of question-answering forums, users and QA boards constitute a heterogeneous network whose nodes are users/boards and edges are authorship. Discovering both user communities and board communities and finding correlations between them will clarify their characteristics. This paper describes an attempt for analyzing heterogeneous social networks obtained from Yahoo! Chiebukuro (Japanese Yahoo! Answers). A new measurement for the correlation between user communities and board communities is defined, and characteristics of discovered communities are analyzed using it.
Subjects: 12. Machine Learning and Discovery; 1.10 Information Retrieval
Submitted: Feb 15, 2008