A Survey of Prediction Using Social Media

《A Survey of Prediction Using Social Media》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ASurveyofPredictionUsingSocialMediaShengYuandSubhashKakDepartmentofComputerScience,OklahomaStateUniversityStillwater,Oklahoma,U.S.A.74078{yshe,subhashk}@cs.okstate.eduAbstract:Socialmediacomprisesinteractiveapplicationsandplatformsforcreating,sharingandexchangeofuser-ge

2、neratedcontents.Thepasttenyearshavebroughthugegrowthinsocialmedia,especiallyonlinesocialnetworkingservices,anditischangingourwaystoorganizeandcommunicate.Itaggregatesopinionsandfeelingsofdiversegroupsofpeopleatlowcost.Miningtheattributesandcontentsofsocialmediagivesusan

3、opportunitytodiscoversocialstructurecharacteristics,analyzeactionpatternsqualitativelyandquantitatively,andsometimestheabilitytopredictfuturehumanrelatedevents.Inthispaper,wefirstlydiscusstherealmswhichcanbepredictedwithcurrentsocialmedia,thenoverviewavailablepredictors

4、andtechniquesofprediction,andfinallydiscusschallengesandpossiblefuturedirections.Keywords:socialmedia,socialnetwork,socialnetworkingservice,user-generatedcontents,prediction1.IntroductionSocialmediaareplatformsthatallowcommonpersonstocreateandpublishcontents.Twoworldwid

5、epopularsocialmediawebsites,TwitterandFacebook,demonstrateitsexplosivegrowthandprofoundinfluence.BothTwitterandFacebookareinthetop10most-visitedwebsitesintheworldaccordingtoAlexaranking[1].Facebookhasmorethan800millionactiveusers[2],andbyMarch2011,onTwitter,therewereabo

6、ut140millioninformationpiecescreatedandtransferreddaily[3].Thereisotherspecializedsocialmediathatarefocusedonentertainment,sports,financeandpolitics.Sincetherearemanyuserssharingtheiropinionsandexperiencesviasocialmedia,thereisaggregationofpersonalwisdomanddifferentview

7、points.Suchaggregationhaslimitationsasviewpointsaresubjecttochangewithtime.Inasensethesocialmediapredictionproblemisparalleledbypredictionoffinancialtimeseriesbasedonpasthistory,whichhasitsusesintrading.Ingeneral,ifextractedandanalyzedproperly,thedataonsocialmediacanlea

8、dtousefulpredictionsofcertainhumanrelatedevents.Suchpredictionhasgreatbenefitsinmanyrealms,suchasfinance,produ