资源描述：
《基于多目标决策的微博用户影响力评价算法-研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、华中科技大学硕士学位论文AbstractWeibohasthecharacteristicsofhightransmissionspeed,extensivesource,andmulti-angle,asaformofsocialmedia.Ithasbecomethemainchannelofinformationexchangeandinformationsharinginpeople’sdailylives,andhasattractedwideattentionofscholarsbothathomeand
2、abroad.Ithasgreatsignificanceinuserrecommendation,informationdiffusion,publicopinionmonitoring,directedmarketingandsoon.First,basedontheanalysisofSinaWeibo’sinformationdiffusioncharacteristics,weconstructaWeibonetworkmodel,andtheWeibonetworkisdividedintotwonetw
3、orks,whichareuserrelationnetworkandtheblogdiffusionnetwork.Second,toeffectivelyavoidtheinfluenceof“zombiefans”andpreventusersfromforwardingorcommentingtheirownblogstoimproveinfluence,thispaperextractsfourindicatorsforevaluatinguserinfluence,whicharetheLeaderRan
4、kinfluence,theaverageretweetednumber,theaveragecommentednumber,andtheaveragelikenumber,fromperspectivesoffollowingrelationshipbetweenusersandblogdiffusion.Basedonthefourindicatorsforevaluatinguserinfluence,weusetheSkylinemethodinthemulti-criteriadecisionmaking.
5、Thenweconstructauserinfluenceevaluationmodel,andproposeaninfluenceevaluationalgorithmnamedasWeiboLeaderRank.Toinvestigatetheeffectivenessofthealgorithm,wedesignandimplementtheSinaWeibodatacollectionsystembyusingwebcrawlertechnique.Becausethewebserverwillredirec
6、taccessrequestandprohibitusertoaccesswhenitdetectsabnormalaccessrequest,whichwillaffectthespeedofthedatacollection,weusemulti-account,simulationloggingandmulti-threadcollection.Thethreadrequestsdatausinganonymousproxyserver,andchangesdynamicallytorequestHTTPhea
7、derinformation.What’smore,weaddanabnormallydetectionmoduletodetecttheabnormalsituationintimeandtakethecorrespondingactions.Wetrytoimitatethenormaluseraccessbehaviortoimprovetheefficiencyofthedatacollection.WeiboLeaderRandistestedonthecollectedlargescaledatasets
8、.TheexperimentalresultsshowthatWeiboLeaderRandperformsbetterthanotherrepresentativeuserinfluenceevaluationalgorithms.Theruntimeincreaseslinearlywithanyincreaseindatasize.Wei
