One Plus One Makes Three (for Social Networks) 英文参考文献.docVIP

OnePlusOneMakesThree(forSocialNetworks) ′ Emo¨ke-AgnesHorva′t1,MichaelHanselmann2,FredA.Hamprecht2,3,KatharinaA.Zweig1,3 * 1InterdisciplinaryCenterforScientificComputing(IWR),UniversityofHeidelberg,Heidelberg,Germany,2HeidelbergCollaboratoryforImageProcessing(HCI),University ofHeidelberg,Heidelberg,Germany,3MarsiliusKolleg,UniversityofHeidelberg,Heidelberg,Germany Abstract Membersofsocialnetworkplatformsoftenchoosetorevealprivateinformation,andthussacrificesomeoftheirprivacy,in exchange for the manifold opportunities and amenities offered by such platforms. In this article, we show that the seeminglyinnocuouscombinationofknowledgeofconfirmedcontactsbetweenmembersontheonehandandtheiremail contacts to non-members on the other hand provides enough information to deduce a substantial proportion of relationships between non-members. Using machine learning we achieve an area under the (receiver operating characteristic) curve (AUC) of at least 0:85 for predicting whether two non-members known by the same member are connectedornot,evenforconservativeestimatesoftheoverallproportionofmembers,andtheproportionofmembers disclosingtheircontacts. Citation: Horva′t E-A′,Hanselmann M,HamprechtFA, ZweigKA(2012) OnePlusOne MakesThree (forSocial Networks).PLoSONE 7(4):e34740. doi:10.1371/ journal.pone.0034740 Editor:SergioGo′mez,UniversitatRoviraiVirgili,Spain ReceivedNovember15,2011;AcceptedMarch5,2012;PublishedApril6,2012 Copyright: ? 2012 Horva′t et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. Funding: EAH and KAZ are supported by the Heidelberg Graduate School of Mathematical and Computational Methods for the Sciences (http://www. mathcomp.uni-heidelberg.de/),UniversityofHeidelberg,Germany,whichisfundedbytheGermanExcellenceInitiative(GSC220).FAHandKAZweresupported byafellowshipoftheMarsiliusKolleg(