基于自適應(yīng)遺傳算法的橢圓聚類方法研究.pdf

1、中北大學(xué)碩士學(xué)位論文基于自適應(yīng)遺傳算法的橢圓聚類方法研究姓名：任建斌申請學(xué)位級別：碩士專業(yè)：應(yīng)用數(shù)學(xué)指導(dǎo)教師：楊明20090401中北大學(xué)學(xué)位論文StudyofEllipticRingshapedClasteringBasedonAdaptiveGeneticAlgorithmAbstractClusteringanalysis，whichisonetypeofmulti—variablestatisicanalysisandansig

2、nificantbranchofunsupervisedpatternrecognition，isdefinedasthestudyandprocessingofthecharacterizationofagivensetofdatawithmathematicapproachAccordingtocertaincriterion，thistechniquepartitionsasampleclusterwithoutgroupmark

3、edintosomegroups，SOthatsimilarsamplesareassignedtoonegroupaspossibleasmuch，whiledissimilartodifferentonesGeneticalgorithm(GA)derivedfromthecomputersimulatedstudyofbiosystem，whichisonekindofcontemporaryoptimizationtechniq

4、ues，andfeaturesfunctionssuchastheoverallandeffectiveparalleloptimization，hi曲robustnessandgeneralityandnodemandforspecificinformationofproblems，orelseDisciplineslikemathrnetics，physics，biologyandcomputerscienceandetcareal

5、lembodiedinthealgorithm，whichshedsnewlightonthesolutionofcomplicatedproblemsInthepresentstudythegeneticalgorithmiscreativelyemployedtotheoptimizationofclusteringobjectivefunctionTheapplicationofAdaptiveGeneticAlgorithm(A

6、GA)tothedetectionofellipticringshapedclustersuccessfullypresentsconvergenceintolocalminimawhichmightresultfromthetraditionalsimpleclusteringmethodThestudyresultprovesthatthenewlyconstructedalgorithmisveryrobustThedissert

7、ationisdividedintofourchapters，withcontentsasfollows：Chapter1istheintroduction，whichintroducesthefundamentalissuesofclusteranalysisandthenecessityandrationalityoftheapplicationofGAtoclusteringobjectivefunctionoptimizatio

8、naswellChapter2looksbackonthedevelopmentalcourseofclusteringalgorithm，whichmainlycentersontheintroductionandderivationoftheFuzzyC—meansAlgorithm(FCM)anddisscussesrespectivelythestrongandweakpointsofthetraditionalclusteri