基于分位數(shù)據(jù)深度的非參數(shù)多變量管制圖.pdf

1、華中科技大學(xué)碩士學(xué)位論文基于分位數(shù)據(jù)深度的非參數(shù)多變量管制圖姓名：王雪申請學(xué)位級別：碩士專業(yè)：概率論與數(shù)理統(tǒng)計指導(dǎo)教師：劉次華20060408IIAbstractAbstractControltwasfirstdevelopedbyShewhartintheyear1931tomonitcontrolamanufacturingprocess.Itisanimptantmethodinstatisticprocesscontrol.Co

2、ntroltsfmanufacturingprocessesareoftenunivariateparametric.Inrealwldhowevermultivariatequalityacteristicsratherthanunivariatearetheonepeopleoftenencounter.Meanwhiletheassumptionofnmalityespeciallythemultivariatenmalityis

3、notalwaysvalid.Whenthenonnmalityisatpresencetheperfmancesoftheseparametriccontroltsaresignificantlyaffected.Thusanonparametricmultivariatecontroltisverydesirable.Whilesomenonparametricmultivariatetshavebeenconstructedbas

4、edonsimplicialdatadepthintroducedbyLiuin1990thisdatadepthisfsomesymmetricdistributionstheamountofcomputationinvolvedinishuge.Thisthesisproposesthequantiledatadepththatcanbeappliedtosomeasymmetricdistributions.Fromthisnew

5、datadepthsomestatisticsthatfollowsomecertaindistributionsarederived.ItisthenreducedtoconstructcommonShewharttCUSUMtEWMAtwiththesestatistics.Thussomenonparametricmultivariatetsarethengiven.Inthisthesisweprovetheconsistenc

6、yofquantiledatadepthputfwardabivariateasymmetricdistributionskewedunifmnmaldistrubutioncomparetheperfmancesofquantiledepthbasedtswithsomeothertswhentheunderlyingdistributionsarebivariatenmalbivariatetofthreedegreesoffree