信息數(shù)據(jù)的數(shù)學(xué)處理方法.pdf

1、上海大學(xué)碩士學(xué)位論文信息數(shù)據(jù)的數(shù)學(xué)處理方法姓名：錢舒申請學(xué)位級別：碩士專業(yè)：計(jì)算數(shù)學(xué)指導(dǎo)教師：許夢杰200301012003年h海大學(xué)碩士學(xué)位論文IiAbstractThisthesismainlystudiestheapplicationofwaveletsinprocessingtheinformationdataoftwodifferentfieldsmetallurgyandfinance，whichshowsthatwavel

2、ets，asanewanddevelopingmathematicalmethods，isvaluableinprocessingtheinformationdataInthefirstplacetheprojectofsolvingthecoefficientsoftheSELFSReMmodelsponsoredbythenationalfundisintroducedIn也efieldofthermodynamicsSELFSRe

3、McanbeusedwidelyforthemostliquidalloywhichisbetterthanthepopularWagnermethodButthemodelcontainsmanyunknowncoefficientsMoreoverforthelimitofexperimentalconditionssuchastheenorofequipments、theelTorofeyes，theexperimentalinf

4、ormationdatacontainerrorsinevitablySotheprojectofsolvingthemodelcoefficientcanbetransformedintoaproblemofsolvingthelinearillposedequationsHowtosolvetheillposedequationsmoreaccuratelyisstilladifficultprobleminthisfieldInt

5、histhesisdiscretewaveletsandMRAayeadoptedtoprocesstheexperimentaldatafirstlyinorderteducethenoiseofdataThentheimplicititerativemethodisappliedtosolvetheill—posedequationsBymeitllsofthenewwaywegetthesarisfyingresultsTheco

6、mputationalexamplesshowthatthisnewwayCanreducethecomputationaltime，improvetheaccuracyandmaketheresultssatisfythephysiealmeaning，TheapplicationofwaveletsinanalyzingstockmarketdataisintroducedinthesecondplaceStockpricesare

7、callednon—stationarytimeseriesWecanuseaLipschitzexponenttoShowitssingularityIn1980sMendalbrotprovedthatthechangeofstockpriceisfractalSecentlytheresearcherstakestockpriceaSonedimensiontimesignalandapplywaveletsinanalyzing

8、thestockdata[7】and【8】introduceanewwaytoanalyzetimeseriesAtfirsttheyestablishtimeseriesmodelbydiscretewavelettransformandMRAThenregressionisusedtoanalyzethedataButthewaveletsfunctiontheyselectedisofashortcomingthatitcan’t

9、containtwomeritsofsmoothqualityandconvergenceTheyselectedthesmoothandnoncompactlysupportedwaveletsfunctionforaccuracywhichmadetheconvergentspeedslowerandbringmuchdifficultytocomputeSoweadopthighspeedsamplingtheoremtoover

10、cometheshortcoming[9】introduceshowtemovetheup—and—downofstockdatacausedbyaccidentalfactorsbymeansofcontinuouswavelettransformforpredictingthetrendofstock塒oe，ButtheeffectisnetverygoodbecauseGaussfunctionasthepolishcellisn

11、’tcompactlysupportedWetakeanewband—limitedmotherwaveletfunctionbuiltinthefirstchapterasthepolishfunctiontofilterthestockmarketdataTheeffectbecomesbetterandthefractalqualitybecomesmoreobvious，whichclarifiesthatstockmarket