Designation:D7915−14AnAmericanNationalStandardStandardPracticeforApplicationofGeneralizedExtremeStudentizedDeviate(GESD)TechniquetoSimultaneouslyIdentifyMultipleOutliersinaDataSet1ThisstandardisissuedunderthefixeddesignationD7915;thenumberimmediatelyfollowingthedesignationindicatestheyearoforiginaladoptionor,inthecaseofrevision,theyearoflastrevision.Anumberinparenthesesindicatestheyearoflastreapproval.Asuperscriptepsilon(´)indicatesaneditorialchangesincethelastrevisionorreapproval.1.Scope1.1ThispracticeprovidesastepbystepprocedurefortheapplicationoftheGeneralizedExtremeStudentizedDeviate(GESD)Many-OutlierProceduretosimultaneouslyidentifymultipleoutliersinadataset.(SeeBibliography.)1.2Thispracticeisapplicabletoadatasetcomprisingobservationsthatisrepresentedonacontinuousnumericalscale.1.3Thispracticeisapplicabletoadatasetcomprisingaminimumofsixobservations.1.4Thispracticeisapplicabletoadatasetwherethenormal(Gaussian)modelisreasonablyadequateforthedistributionalrepresentationoftheobservationsinthedataset.1.5Theprobabilityoffalseidentificationofoutliersasso-ciatedwiththedecisioncriteriasetbythispracticeis0.01.1.6ItisrecommendedthattheexecutionofthispracticebeconductedundertheguidanceofpersonnelfamiliarwiththestatisticalprinciplesandassumptionsassociatedwiththeGESDtechnique.1.7Thisstandarddoesnotpurporttoaddressallofthesafetyconcerns,ifany,associatedwithitsuse.Itistheresponsibilityoftheuserofthisstandardtoestablishappro-priatesafetyandhealthpracticesanddeterminetheapplica-bilityofregulatorylimitationspriortouse.2.Terminology2.1DefinitionsofTermsSpecifictoThisStandard:2.1.1outlier,n—anobservation(orasubsetofobservations)whichappearstobeinconsistentwiththeremainderofthedataset.3.SignificanceandUse3.1TheGESDprocedurecanbeusedtosimultaneouslyidentifyuptoapre-determinednumberofoutliers(r)inadataset,withouthavingtopre-examinethedatasetandmakeaprioridecisionsastothelocationandnumberofpotentialoutliers.3.2TheGESDprocedureisrobusttomasking.Maskingdescribesthephenomenonwhere...
