A Comprehensive Survey of Clustering Algorithms

A Comprehensive Survey of Clustering Algorithms

ID:41100587

大小:570.02 KB

页数:29页

时间:2019-08-16

A Comprehensive Survey of Clustering Algorithms_第1页
A Comprehensive Survey of Clustering Algorithms_第2页
A Comprehensive Survey of Clustering Algorithms_第3页
A Comprehensive Survey of Clustering Algorithms_第4页
A Comprehensive Survey of Clustering Algorithms_第5页
资源描述:

《A Comprehensive Survey of Clustering Algorithms》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、Ann.Data.Sci.DOI10.1007/s40745-015-0040-1AComprehensiveSurveyofClusteringAlgorithmsDongkuanXu1,2·YingjieTian2,3Received:25May2015/Revised:18July2015/Accepted:31July2015©Springer-VerlagBerlinHeidelberg2015AbstractDataanalysisisusedasacommonmethodinmodernscienceresearch,whichisacrossc

2、ommunicationscience,computerscienceandbiologyscience.Clus-tering,asthebasiccompositionofdataanalysis,playsasignificantrole.Ononehand,manytoolsforclusteranalysishavebeencreated,alongwiththeinformationincreaseandsubjectintersection.Ontheotherhand,eachclusteringalgorithmhasitsownstrengt

3、hsandweaknesses,duetothecomplexityofinformation.Inthisreviewpaper,webeginatthedefinitionofclustering,takethebasicelementsinvolvedinthecluster-ingprocess,suchasthedistanceorsimilaritymeasurementandevaluationindicators,intoconsideration,andanalyzetheclusteringalgorithmsfromtwoperspecti

4、ves,thetraditionalonesandthemodernones.AllthediscussedclusteringalgorithmswillbecomparedindetailandcomprehensivelyshowninAppendixTable22.KeywordsClustering·Clusteringalgorithm·Clusteringanalysis·Survey·UnsupervisedlearningBYingjieTiantyj@ucas.ac.cnDongkuanXuxudongkuan14@mails.ucas.a

5、c.cn1SchoolofMathematicalSciences,UniversityofChineseAcademyofSciences,Beijing100049,China2ResearchCenteronFictitiousEconomy&DataScience,ChineseAcademyofSciences,Beijing100190,China3KeyLaboratoryofBigDataMiningandKnowledgeManagement,ChineseAcademyofSciences,Beijing100190,China123Ann

6、.Data.Sci.1IntroductionClustering,consideredasthemostimportantquestionofunsupervisedlearning,dealswiththedatastructurepartitioninunknownareaandisthebasisforfurtherlearning.Thecompletedefinitionforclustering,however,isn’tcometoanagreement,andaclassiconeisdescribedasfollows[1]:(1)Insta

7、nces,inthesamecluster,mustbesimilarasmuchaspossible;(2)Instances,inthedifferentclusters,mustbedifferentasmuchaspossible;(3)Measurementforsimilarityanddissimilaritymustbeclearandhavethepracticalmeaning;Thestandardprocessofclusteringcanbedividedintothefollowingseveralsteps[2]:(1)Featu

8、reextractionandsele

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。