资源描述:
《聚类分析:基本概念及算法 毕业论文外文翻译》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、21北京交通大学毕业设计(论文)第页ClusterAnalysis:BasicConceptsandAlgorithmsClusteranalysisdividesdataintogroups(clusters)thataremeaningful,useful,orboth.Ifmeaningfulgroupsarethegoal,thentheclustersshouldcapturethenaturalstructureofthedata.Insomecases,however,clusteranalys
2、isisonlyausefulstartingpointforotherpurposes,suchasdatasummarization.Whetherforunderstandingorutility,clusteranalysishaslongplayedanimportantroleinawidevarietyoffields:psychologyandothersocialsciences,biology,statistics,patternrecognition,informationretrieva
3、l,machinelearning,anddatamining.Therehavebeenmanyapplicationsofclusteranalysistopracticalproblems.Weprovidesomespecificexamples,organizedbywhetherthepurposeoftheclusteringisunderstandingorutility.ClusteringforUnderstandingClasses,orconceptuallymeaningfulgrou
4、psofobjectsthatsharecommoncharacteristics,playanimportantroleinhowpeopleanalyzeanddescribetheworld.Indeed,humanbeingsareskilledatdividingobjectsintogroups(clustering)andassigningparticularobjectstothesegroups(classification).Forexample,evenrelativelyyoungchi
5、ldrencanquicklylabeltheobjectsinaphotographasbuildings,vehicles,people,animals,plants,etc.Inthecontextofunderstandingdata,clustersarepotentialclassesandclusteranalysisisthestudyoftechniquesforautomaticallyfindingclasses.Thefollowingaresomeexamples:21北京交通大学毕业
6、设计(论文)第页Biology.Biologistshavespentmanyyearscreatingataxonomy(hierarchicalclassification)ofalllivingthings:kingdom,phylum,class,order,family,genus,andspecies.Thus,itisperhapsnotsurprisingthatmuchoftheearlyworkinclusteranalysissoughttocreateadisciplineofmathe
7、maticaltaxonomythatcouldautomaticallyfindsuchclassificationstructures.Morerecently,biologistshaveappliedclusteringtoanalyzethelargeamountsofgeneticinformationthatarenowavailable.Forexample,clusteringhasbeenusedtofindgroupsofgenesthathavesimilarfunctions.•Infor
8、mationRetrieval.TheWorldWideWebconsistsofbillionsofWebpages,andtheresultsofaquerytoasearchenginecanreturnthousandsofpages.Clusteringcanbeusedtogroupthesesearchresultsintoasmallnumberofclusters,eachofwh