欢迎来到天天文库
浏览记录
ID:34004185
大小:911.32 KB
页数:43页
时间:2019-03-03
《基于mmsb模型的微博网络社区发现的研究》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、AbstractAbstractCommunitydetectionmainlybasedonnodelinkdensenessincommunitiesandlinksparsenesscharacter!sticsbetweencommunities,andwiththehelpofthenetworkaveragepathlength,clusteringcoefficientanddegreedistributionofdepictingthenetworkstructureofattributest
2、odifferentiatecommunity.Thesemethodscan'tfindoverlappingcommunityanditscomputationalcomplexitywillbecomelargerbecauseoftheincreasingofthenetworknodes.Asaresultofthehighnumberofthenetworknodesandthecomplexrelationshipbetweenthenodes,thetraditionalmethodhasli
3、mitationsforMicro-blognetwork.MMSBmodelisakindbasedBaesaetwrkadallwsdesbeltltpledelsthect.Althoughthemodelcandealwithlarge-scalenetworknodes,andthecomputingcomplexityisrelativelylow,butthemodelissuitableforundirectednetwork.BasedontheMMSBmodelandcombinedthe
4、characteristicsofMicro-blog,byimprovingtheoldmodel,IputforwardanewcommunitydetectionmodelWB-MMSBforMicro-blog.ConsideringthenodesofMicro-blogNetworkhavesingledirectionrelations,WB-MMSBisptrwardrctdetectwhchsesdrectededestebdthedrectrelationsofnodes,fromtwoa
5、spectslink-inandlink-outtoquantifythecommunitymembershipofnodes,usesexponentialfamilydistributionandmean-fieldvariationalinferencemethodtoinferencetherepresentationsofvariablesinthismodel,andusesSVIalgorithmtocomputerelatingparameters.ExperimentsadoptSina-W
6、eibodatasetandNMItotesttheperraceWB-MMSB.TheresltsdcatethatthectdetectabilityofWB-MMSBmodelisbetterthanaMMSBmodel,theconvergencerateofWB-MMSBmodelisfasterthanaMMSBmodel.KeywordsMicro-blogNetworkCommunitydetectionMixedmembershipstochasticblockmodelOverlappin
7、gcommunitiesSVIIIWX´NH11J&.............................................................................................................11.1Ẇ(OP#RS............................................................................................11.2Ẇ(äT...........
8、.............................................................................................21.2.1Ẇ(äT.....................................................................................21.2.2âãKäẆ
此文档下载收益归作者所有