RBF的改進及在中國轎車市場預(yù)測中的應(yīng)用研究.pdf

1、碩士學(xué)位論文RBF的改進及在中國轎車市場預(yù)測中的應(yīng)用研究ResearchonImprovementandApplicationofRBFinForecastingCarMarket怍g姓名：——上L—簋—玉L學(xué)科、n——亡』L璺丑聾學(xué)號———!盟Ⅱ地!一一指齄∞——點』生型—煎控—完成H期：——』鯉￡主塑上——一九連“工大學(xué)DalianUaivcrsJtyofTechnology尹德武：RBF的改進及在中國轎車市場預(yù)測中的應(yīng)用研究Res

2、earchonImprovementandApplicationofRBFinForecastingCarMarketAbstractWiththedevelopmentofeconomyandtheimprovementofpeople’slivingstandardstheconsumptionofcarhasbecomeahotspotpeoplepayattentiontoItisthekeypointthatcarsenter

3、intothecommonpeople’Sfamilies，accordingtothedevelopmentcourseofcarindustryMeanwhile，itmakescarindustryimportantinstrumentsinwhichnationstartupseconomyandstimulatesdomesticdemandTherefore，itisnecessarytoforecastthedemandc

4、arinthefulureItisSOdifficultthattheresearchershavespentlotsofeffortsinstudytasksAtpresentthereare200forecastmethods，30ofwhicharewidelyusedandmorethan10ofwhichareusedcommonlyThisdataincreasesconstantlywiththeadvanceoftime

5、Variouscorrelativefactorshavingeffectonthedemandofcarmarketarenotsimplelinearrelationbutcomplexnon1inearrelationThetraditionalforecastmethodsfacedifficultieswhentheyresolvetheseproblemsItmakestheforecastresultsnotsatisfi

6、edInOrdertoobtainmoresatisfiedresultweconstantlyattempttoapplynewscientifictheoriesandtechnologiesteasonableforecastinforecastingthedemandofcarmarketNervenetworkmethodisascience，springingupinrecentyearsWiththedevelopment

7、ofcomputerscienceitsapplicationhasbecomemoreandmorewideHoweverthenervenetworkmethodiSstillattheelementarystageofexplorationandpractice，wefindoutthatradialbasisnetworkhasbetterforecastcharacteristicbythedoingresearchonner

8、venetworkmethodFirStofallthispaperutilizesthetraditionalmethodstoforecastandanalyzesthereasonsfortheerrorsfromtheforecastresultsandtodiscoverthecomplexrelationsbetweenvariouscorrelativefactorsandeardemandquantumThen，usin

9、gthethoughtofequaldimensionalityvectorsimprovestheradialbasisnetworkWefindoutthatRBFnetworkcanresolvetheproblemsthattraditionalmethodscannotresolve，throughcomparingRBFnetworkwithtraditionalmethodsinforecastingthedemandre

10、sultsofcarmarketintheseveralyears’futureRBFnetworkutilizestheprincipleofartificialintelligenceandmultilayerinformationtransferredtoestablishthecomplexrelationbetweenafter—variableandbefore—variableandjthasobviousadvantag