comparison and validation of community structures in complex networks

comparison and validation of community structures in complex networks

ID:34149042

大小:251.63 KB

页数:25页

时间:2019-03-03

comparison and validation of community structures in complex networks_第1页
comparison and validation of community structures in complex networks_第2页
comparison and validation of community structures in complex networks_第3页
comparison and validation of community structures in complex networks_第4页
comparison and validation of community structures in complex networks_第5页
资源描述:

《comparison and validation of community structures in complex networks》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、ComparisonandvalidationofcommunitystructuresincomplexnetworksMikaGustafsson,MichaelHornquist¨∗,AnnaLombardiDivisionofPhysicsandElectronics,DepartmentofScienceandTechnology,Linkoping¨University,SE-60174Norrkoping,Sweden¨AbstractTheissueofpartitioninganetwo

2、rkintocommunitieshasattractedagreatdealofatten-tionrecently.Mostauthorsseemtoequatethisissuewiththeoneoffindingthemaximumvalueofthemodularity,asdefinedbyNewman.SincetheproblemformulatedthiswayisNP-hard,mostefforthasgoneintotheconstructionofsearchalgorithms,

3、andlesstothequestionofothermeasuresofcommunitystructures,similaritiesbetweenvariouspartition-ingsandthevalidationwithrespecttoexternalinformation.Hereweconcentrateonaclassofcomputergeneratednetworksandonthreewell-studiedrealnetworkswhichconstituteabench-m

4、arkfornetworkstudies;thekarateclub,theUScollegefootballteamsandagenenetworkofyeast.Weutilizesomestandardwaysofclusteringdata(originallynotdesignedforfindingcommunitystructuresinnetworks)andshowthattheseclassicalmethodssometimesoutperformthenewerones.Wedisc

5、ussvar-iousmeasuresofthestrengthofthemodularstructure,andshowbyexamplesfeaturesanddrawbacks.Further,wecomparedifferentpartitionsbyapplyingsomegraph-theoreticconceptsofdistance,whichindicatethatoneofthequalitymeasuresofthedegreeofmodu-laritycorrespondsquit

6、ewellwiththedistancefromthetruepartition.Finally,weintroduceawaytovalidatethepartitioningswithrespecttoexternaldatawhenthenodesareclassi-fiedbutthenetworkstructureisunknown.Thisisherepossiblesinceweknoweverythingofthecomputergeneratednetworks,aswellasthehi

7、storicalanswertohowthekarateclubandthefootballteamsarepartitionedinreality.Thepartitioningofthegenenetworkisval-idatedbyuseoftheGeneOntologydatabase,whereweshowthatacommunityingeneralcorrespondstoabiologicalprocess.arXiv:physics/0601057v1[physics.soc-ph]1

8、0Jan2006Keywords:network,community,validation,distancemeasure,hierarchicalclustering,K-means,GOPACS:89.75.Fb,89.75.Hc,87.16.Yc02.10.Ox∗CorrespondingauthorEmailaddress:[mikgu,micho,annlo]@itn.liu.se(MikaGustafsson,Mi

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。