overview on techniques in cluster analysis

overview on techniques in cluster analysis

ID:7348919

大小:704.84 KB

页数:27页

时间:2018-02-12

overview on techniques in cluster analysis_第1页
overview on techniques in cluster analysis_第2页
overview on techniques in cluster analysis_第3页
overview on techniques in cluster analysis_第4页
overview on techniques in cluster analysis_第5页
资源描述:

《overview on techniques in cluster analysis》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库

1、Chapter5OverviewonTechniquesinClusterAnalysisItziarFradesandRuneMatthiesenAbstractClusteringistheunsupervised,semisupervised,andsupervisedclassificationofpatternsintogroups.Theclusteringproblemhasbeenaddressedinmanycontextsanddisciplines.Clusteranalysisencompassesdifferentmethodsandalgorithmsforg

2、roupingobjectsofsimilarkindsintorespectivecategories.Inthischapter,wedescribeanumberofmethodsandalgorithmsforclusteranalysisinastepwiseframework.Thestepsofatypicalclusteringanalysisprocessincludesequentiallypatternrepresentation,thechoiceofthesimilaritymeasure,thechoiceoftheclusteringalgorithm,t

3、heassessmentoftheoutput,andtherepresentationoftheclusters.Keywords:Clusteringalgorithm,featureselection,featureextraction,similaritymeasure,clustertendency,clustervalidity,clusterstability,relevancenetworks,dendrogram.1.Introduction1.1.TheImportanceClusteringisoneofthemostusefultasksinthedatamin

4、ingpro-ofClusteringcessfordiscoveringgroupsandidentifyingnewinterestingpat-ternsintheunderlyingdata.Clusteringalgorithmspartitiondataobjectsintosubsets(clusters)basedonsimilarityordissimilar-ity.Patternswithinavalidclusteraremoresimilartoeachotherthantheyaretoapatternbelongingtoadifferentcluster

5、.Theclusteringprocessisanunsupervised,semisupervised,orsuper-visedmethod.Sinceunsupervisedclusteralgorithmsdonotusepredefinedclasslabelsorexamplesthatwouldindicategroupingpropertiesinthedataset,itistheidealmethodforidentifyingnewpatternsindata.Unsupervisedclusteringisalsofrequentlyusedincombinati

6、onwithothersupervisedclassificationalgorithmssinceithasthepotentialtodetectincorrectclasslabels,outliers,errors,bias,andbadexperimentaldesigns.R.Matthiesen(ed.),BioinformaticsMethodsinClinicalResearch,MethodsinMolecularBiology593,DOI10.1007/978-1-60327-194-35,©HumanaPress,apartofSpringerScience+B

7、usinessMedia,LLC20108182FradesandMatthiesen(15)(15)Relevancenetworks(9)Displayingtheassessmentoftheuncertaintyinhier-archicalclusteranalysis(27)RepresentationofclustersGraphsPartitionsClassificationtreesDendrogram(7,8)Externa

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。