作者: Howard T. Welser , Danyel Fisher , Marc A. Smith , Eric Gleave
DOI:
关键词: Ordinary least squares 、 Taxonomy (general) 、 Online discussion 、 Psychology 、 World Wide Web 、 Visualization methods 、 Identification (information) 、 Regression analysis
摘要: Social roles in online discussion forums can be described by patterned characteristics of communication between network members which we conceive as 'structural signatures.' This paper uses visualization methods to reveal these structural signatures and regression analysis confirm the relationship their associated Usenet newsgroups. Our focuses on distinguishing one role from others, "answer people." Answer people are individuals whose dominant behavior is respond questions posed other users. We found that answer predominantly contribute or a few messages discussions initiated disproportionately tied relative isolates, have intense ties triangles local networks. OLS shows strongly correlated with and, combination, provide predictive model for identifying (R =.72). To conclude, consider strategies further improving identification settings how development taxonomy author types could extended newsgroups particular systems general. 2