《Metrics in Sharing Video Relationship-Networks》.pdf

ChineseJournalofElectronics Vo1．23，No．3，July2014 M etricsin SharingVideoRelationship—Networks术 GAOWen．QIANYaguan andWUChunming (1．CollegeofComputerScience＆Technology，ZhejiangUniversity，Hangzhou310027,China) (2．CollegeofScience，Zh~iangUniversityofScienceandTechnology，Hangzhou310023，China) Abstract—— AtPresentthevideo sharing service has the topologicalcharacteristicsofSVRNshavebeen studied gradually becam eoneofthe Internetkillerapplications．It and some “small—world”featureshavebeenrevealedt5』．How— is crucial to the service providers that understand the ever．morein—depthexploitation isstillneededtorevealother intrinsic nature of Sharing video relationship．networks facetsoftheSVRNstopologythroughextensiveanalysiswith fSVRNs) formed by video related list．ThisPaPer in— differentmetrics．Thisresearchcouldenableustoquantifyand vestigates the relati0nships between sharing videos us— ing com plex networkstheory．N ew m etricsare introdueed modelthecorrelationamongthevideosinSVRNs．andunder— to m odel and characterize the SVRN from topol0gical standtheirinteractivePatterns．In addition．itisparamount view ．Through extensive study on Y-0uTube．a set ofkey importantforserviceandcontentproviderstooptimallyplan findingsare revealed．Theresultsshow thatthenode in— andhigh—efficiently delivertheirservicestotheendusers． degree ofSV RNs follows the power—law distribution and In fact，therelationshipsbetween sharing videoscan be itsentropyrem ainsstableovertim e．Thetopologyalso ex- modeled as adir