基于證據(jù)理論的模糊kNN分類(lèi)方法及其在故障診斷中的應(yīng)用研究.pdf

1、河北師范大學(xué)碩士學(xué)位論文基于證據(jù)理論的模糊kNN分類(lèi)方法及其在故障診斷中的應(yīng)用研究姓名：杜妮申請(qǐng)學(xué)位級(jí)別：碩士專(zhuān)業(yè)：無(wú)線(xiàn)電物理指導(dǎo)教師：魏志成；呂鋒20120301AbstractWimtherapiddevelopmentofmoderneconomythemodemindustry’Sscaleisbecominglagerandthedegreeofthecomplexityofindustrialprocessishighera

2、ndhigherandasaresult，themodemsystemismoreandmorecomplexFaultdiagnosismethodsbasedondatadrivenarestudiedinthispaperespeciallythemachinelearningmethodsgiventhesituationthattheaccuratemodelofacomplexsystemisveryhardtoobtain

3、whilethehistoricaldataofthesystemundertheconditionsofnormalandallcategoriesoffaultsarerelativelyeasytogetkNNisoneofthemachinelearningmethodswhichiscommonlyusedAsatheoreticalperfectmethod，kNNhastheadvantagesofsimpleandhig

4、hercorrectclassificationratewhilenotrequiringtheprocessoftrainingFromtheabove，kNNanditsapplicationinfaultdiagnosisarefurtherimprovedinthispaperbasedontheresearchonkNNandavarietyofmodifiedkNNmethodsThemainworkandcontribut

5、ionsareasfollows：(1)kNNandthemodifiedkNNmethodsespeciallyFkNNandEkNNarestudiedparticularlyinthispapermeanwhile，theadvantagesandthedisadvantagesofthesemethodsareconcluded(2)AmethodcalledFEkNNispresentedinthispapertooverco

6、metheshortcomingsofkNNandthemodifiedkNNmethodsespeciallyFkNNandEkNNsummarizedaboveInthismethod，twoproblemsthatthedifferencesofthesamplefeaturescan’tberecognizedandtheeffectoffuzzinessthatarousedbythedifferentdistancesbet

7、weenneighborsandthecenterofclassesisnottakenintoaccountaresolved(3)Afundamentalprobleminmulti—classifiersystemishowtocombinethedecisionsmadebytheclassifiersaccordingtosomewayteachconsensusgiventheclassifiers’differentper