Computing Social Communities in Asymmetric Social Media like Twitter

摘要

An asymmetric Social Media source like Twitter is somewhat like a 'Citizen's Band' radio in that it allows people to follow others irrespective of reciprocity: because someone follows me I don't necessarily follow them. It is more challenging to define social network in this form of media since, unlike a 'friend' linkage, a link in this network does not conform to the mutual reciprocity that is normally expected. In this paper, we demonstrate a social network calculation approach that works with both symmetric social media sources like Facebook as well as asymmetric sources like Twitter. We demonstrate the differences between the networks and how the revealed structure can be used to interpret topics of conversation, group influence and other important indicators of social activity.