作者: Lise Getoor , Christopher P. Diehl , Galileo Namata
DOI:
关键词:
摘要: Online communications provide a rich resource for understanding social networks. Information about the actors, and their dynamic roles relationships, can be inferred from both communication content traffic structure. A key component in analysis of online such as email is resolution name references within body message. Name reference relies on context message; message sender recipients’ relationships help to resolve reference. Here we investigate variety approaches which make use network disambiguate references. The serves proxy inferring relationships. These turn us infer likely candidates Our initial findings suggest that simple temporal models effectively For class proposed, performance maximized by exploiting long-term statistics rank candidates.