AFS模糊邏輯在分類器設(shè)計中的應(yīng)用.pdf

1、大連理工大學(xué)碩士學(xué)位論文AFS模糊邏輯在分類器設(shè)計中的應(yīng)用姓名：劉慧申請學(xué)位級別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：劉曉東20081201APS模糊邏輯在分類器設(shè)計中的應(yīng)用TheApplicationofAFSFuzzyLogicintheDesignofClassifierAbstractAlongwiththematureapplicationofdatabasetechniqueandtherapiddevelopmentof

2、theInternettechnology，thequantityofdatainpeople’SdaffyfivesisincreasingrapidlybyexponentialThereforeweneedanewmethodtofindthedeeperrulewhichcanhelpprovidemoreefficientdecisionsupportsincepeoplelifenotsatisfiedwithconvent

3、ionalsearchingandstatisticsanalysismethodanymoreThenmoreandmoreclassificationproblemsappearinourdailyfives，suchasmedicine，agricultureandsecurityfieldeteInordertosolvethepracticalproblemsbettermanymethodshavebeentriedford

4、esigningtheclassifiersuchasthemethodlikefuzzylogic，geneticalgorithm，decisiontreealgorithmsandneuralnetworkalgorithmetewhichhaveachievedsua：esstosomedegree，aremainlyappliedinclassifierdesignSincetheapplicationfieldisbecom

5、ingwiderandwiderthestudyonclassificationproblemhasbrighterfutureandmakethestudymoremeaningfulAccordingtotheshorcomingsofexistingclassifierbasedonAFSfuzzylogic，anewclassifierbasedonAFS(AxiomaticFuzzySet)fuzzylogicWasdesig

6、nedTheclassifierhasseveraladvantages，suchasshortandaccuratedescription，smalltimeconsumption，easiersentencesforbetterunderstandingrelativelyhighaccuracyExperimentalresultsdemonstratedthatahighaccuracyCallbeachievedusingth

7、eproposedclassificationalgorithmbyapplyingonlytheorderrelationsoftheattributes，andthenumericalrepresentationsoftheattributealenotnecessarycompositionsThus，thisalgorithmiswidelyapplicable，andthatitcanretainmoreinformation

8、oftheoriginaldataExperimentsofclassificationweredoneonthreefamousdatasets：wine，irisandWBCD(Wisconsinbreastcancerdata)，whichwereobtainedfiromdatabaseofUCI(UniversityofCaliforniaIrvine)Afteranalyingexperimentresults，wefoun

9、doutsomeproblemsexistedintheAFSLogicalgorithm，suchastheeasymissingproblemincaseofmulti—parameterscombinationsandthatsomeofthedescriptionsweredifficulttounderstandFortheseproblems，modificationandimprovementofalgorithmwere

10、doneandbetterperformanceswereachievedfromthenewexperimentsonfourdatasetsAtlast，throughthecomparewithotheralgorithms，theadvantagesofthisclassificationalgorithmwereshownKeyWords：AFSFuzzyLogic；DataMining；Classification：Fuzz