Influence patterns in topic communities of social media

Kardara, M; Papadakis, G; Papaoikonomou, T; Tserpes, K; Varvarigou, T

dc.contributor.author	Kardara, M	en
dc.contributor.author	Papadakis, G	en
dc.contributor.author	Papaoikonomou, T	en
dc.contributor.author	Tserpes, K	en
dc.contributor.author	Varvarigou, T	en
dc.date.accessioned	2014-03-01T02:53:49Z
dc.date.available	2014-03-01T02:53:49Z
dc.date.issued	2012	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/36478
dc.subject	Social influence	en
dc.subject	Social media	en
dc.subject	Topic Communities	en
dc.subject.other	Aggregate behavior	en
dc.subject.other	Data mining techniques	en
dc.subject.other	Large parts	en
dc.subject.other	On-the-fly	en
dc.subject.other	Representative sample	en
dc.subject.other	Social influence	en
dc.subject.other	Social media	en
dc.subject.other	Targeted marketing	en
dc.subject.other	Semantic Web	en
dc.subject.other	Semantics	en
dc.subject.other	Aggregates	en
dc.title	Influence patterns in topic communities of social media	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1145/2254129.2254144	en
heal.identifier.secondary	http://dx.doi.org/10.1145/2254129.2254144	en
heal.identifier.secondary	10	en
heal.publicationDate	2012	en
heal.abstract	Users of Social Media typically gather into communities on the basis of some common interest. Their interactions inside these on-line communities follow several, interesting patterns. For example, they differ in the level of influence they exert to the rest of the group: some community members are actively involved, affecting a large part of the community with their actions, while the majority comprises plain participants (e.g., information consumers). Identifying users of the former category lies on the focus of interest of many recent works, as they can be employed in a variety of applications, like targeted marketing. In this paper, we build on previous research that examined influencers in the context of a popular Social Media web site, namely Twitter. Unlike existing works that consider its user base as a whole, we focus on communities that are created on-the-fly by people that post messages about a particular topic (i.e., topic communities). We examine a large and representative sample of real-world communities and evaluate to which extent their influential users determine the aggregate behavior of the entire community. To this end, we consider a practical use case: we check whether the community's overall sentiment stems from the aggregate sentiment of this core group. We also examine the dynamics of groups of influencers by assessing the strength of the ties between them. In addition, we identify patterns in the content produced by influencers and the relation between influencers of different communities. Our experiments lead to interesting conclusions that highlight many aspects of in-fluencers' activity inside topic communities; thus, they form the basis for intelligent, data mining techniques that can automatically discover influencers in the context of a community. Copyright 2012 ACM.	en
heal.journalName	ACM International Conference Proceeding Series	en
dc.identifier.doi	10.1145/2254129.2254144	en