资源描述：
《Influential Observations in Linear Regression.pdf》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、InfluentialObservationsinLinearRegressionAuthor(s):R.DennisCookReviewedwork(s):Source:JournaloftheAmericanStatisticalAssociation,Vol.74,No.365(Mar.,1979),pp.169-174Publishedby:AmericanStatisticalAssociationStableURL:http://www.jstor.org/stable/2286747.Accessed:28/12/201202:18YouruseoftheJSTORar
2、chiveindicatesyouracceptanceoftheTerms&ConditionsofUse,availableat.http://www.jstor.org/page/info/about/policies/terms.jsp.JSTORisanot-for-profitservicethathelpsscholars,researchers,andstudentsdiscover,use,andbuilduponawiderangeofcontentinatrusteddigitalarchive.Weuseinformationtechnologyandtool
3、stoincreaseproductivityandfacilitatenewformsofscholarship.FormoreinformationaboutJSTOR,pleasecontactsupport@jstor.org..AmericanStatisticalAssociationiscollaboratingwithJSTORtodigitize,preserveandextendaccesstoJournaloftheAmericanStatisticalAssociation.http://www.jstor.orgThiscontentdownloadedon
4、Fri,28Dec201202:18:49AMAllusesubjecttoJSTORTermsandConditionsInfluentialObservationsinLinearRegressionR.DENNISCOOK*Characteristicsofobservationswhichcausethemtobeinfluentialdesigntobeinsensitivetooutliers,theviishouldbeinaleastsquaresanalysisareinvestigatedandrelatedtoresidualconstant.variances
5、,residualcorrelations,andtheconvexhulloftheobservedvaluesoftheindependentvariables.ItisshownhowdeletinganAcloseexaminationoftheresidualvariancesor,observationcansubstantiallyalterananalysisbychangingtheequivalently,thevariancesofthepredictedvalues,seemspartialF-tests,thestudentizedresiduals,the
6、residualvariances,thenecessaryintheanalysisofanyexperiment.Apartfromconvexhulloftheindependentvariables,andtheestimatedparame-tervector.Outliersarediscussedbriefly,andanexampleispresented.aproportionalityconstant,thesevariancesaredeter-minedbythedesignpoints.WewillrefertothesmallestKEYWORDS:Del
7、etingobservations;Outliers;PartialF-tests;Residualcorrelations;Studentizedresiduals.convexsetcontainingalldesignpointsastheindependentvariablehull(IVH).1.INTRODUCTIONCook(1977)developedameasurebasedonconfidenceellipsoidsforjudging
