一種新的模糊規(guī)則權(quán)重方法的數(shù)據(jù)分類的研究.pdf

1、大連海事大學(xué)碩士學(xué)位論文一種新的模糊規(guī)則權(quán)重方法的數(shù)據(jù)分類的研究姓名：馮丹申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：應(yīng)用數(shù)學(xué)指導(dǎo)教師：陳剛20100601英文摘要ABSTRACTClassificationisoneof也emostimportantresearchcontentsindataminingItisatechnologythatconstructsclassificationmodelbyanalyzinggivendatasetsandc

2、lassifiessamplesofknownclassbyusingaboveclassificationmodelThetraditionalclassificationmethodsthatdealwithwellbalanceddataCalloftenobtaingoodclassificationperformanceHoweverthereexistsmoreimbalanceddataintherealworldFord

3、ealing、訪也classificationofimbalanceddatathetraditionalclassificationmethodsoftentendtothemajorityclassandleadalowerclassificationaccuracytotheminorityclassThus，itisveryimportanttomakeresearchesontheclassificationmethodsof

4、theimbalanceddataTodifferemtypesofdatasets，thispaperproposesaclassificationalgorithmbasedonfuzzyrules，whichcanadjusttheclassificationaccuracyonthewellbalancedataandtheimbalanceAdataeffectivelyWedesignaclassificationalgor

5、ithmthatincludestheChietalalgorithmandthefuzzyreasoningmodelItintroducesweightingcoefficientsandpatterndistributionfunctiontoimprovethecalculationmethodoffuzzyrulesweightsThisalgorithmnotonlykeepsthepatternmatchingdegree

6、withinclassinuniformdistributionbutalsoenhancesthecontrastofinterclassMoreoveritweaksthegapofwithinclassandenhancesthedifferenceofinterclassSo，classificationaccuracyisimprovedbytherulesweightsadjustedbymeansofweightingco

7、efficientsInparticularweapplytheSMOTEalgorithmtopreprocesstheimbalanceddatawhichleadstobasicbalancebetweentheminorityclassandmajorityclassinquantityBasedonthestudies，weclassifytheimbalanceddatabyusingtheclassificationalg

8、orithmThenweverifythereliabilityoftheclassificationalgorithmonnumericalsimulationaboutthewellbalancedandimbalanceddatawhichhavethedifferentirnbalanceddegreesofUCIdatasetsInadditionwecompareourexperimentalresultswitllothe