资源描述：
《A Fast Model Selection Procedure for Large Families of Models》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、Thisarticlewasdownloadedby:[ShanDongUniversity]On:07May2013,At:06:55Publisher:Taylor&FrancisInformaLtdRegisteredinEnglandandWalesRegisteredNumber:1072954Registeredoffice:MortimerHouse,37-41MortimerStreet,LondonW1T3JH,UKJournaloftheAmericanStatisticalAssociationPublicationdetails,includingi
2、nstructionsforauthorsandsubscriptioninformation:http://amstat.tandfonline.com/loi/uasa20AFastModelSelectionProcedureforLargeFamiliesofModelsabcDavidEdwards&TomášHavránekaNovoResearchInstitute,DK-2880,Bagsvaerd,DenmarkbBiostatisticalInstitute,CopenhagenUniversity,DenmarkcGeneralComputingCen
3、tre,CzechoslovakAcademyofSciences,18207,Prague,CzechoslovakiaPublishedonline:12Mar2012.Tocitethisarticle:DavidEdwards&TomášHavránek(1987):AFastModelSelectionProcedureforLargeFamiliesofModels,JournaloftheAmericanStatisticalAssociation,82:397,205-213Tolinktothisarticle:http://dx.doi.org/10.1
4、080/01621459.1987.10478421PLEASESCROLLDOWNFORARTICLEFulltermsandconditionsofuse:http://amstat.tandfonline.com/page/terms-and-conditionsThisarticlemaybeusedforresearch,teaching,andprivatestudypurposes.Anysubstantialorsystematicreproduction,redistribution,reselling,loan,sub-licensing,systema
5、ticsupply,ordistributioninanyformtoanyoneisexpresslyforbidden.Thepublisherdoesnotgiveanywarrantyexpressorimpliedormakeanyrepresentationthatthecontentswillbecompleteoraccurateoruptodate.Theaccuracyofanyinstructions,formulae,anddrugdosesshouldbeindependentlyverifiedwithprimarysources.Thepubl
6、ishershallnotbeliableforanyloss,actions,claims,proceedings,demand,orcostsordamageswhatsoeverorhowsoevercausedarisingdirectlyorindirectlyinconnectionwithorarisingoutoftheuseofthismaterial.AFastModelSelectionProcedureforLargeFamiliesofModelsDAVIDEDWARDSandTOMASHAVRANEK*~~Anefficientprocedure
7、formodelselectionfromlargefamiliesofmodelsof-fittestformodelsin5.Weassumethatthetestisappliedisdescribed.ItiscloselyrelatedtotheallpossiblemodelsapproachbutforsomeparticularsignificancelevelLY(whichisusuallyisconsiderablyfaster.Itisbasedontwoprinciples:first,i
