ASTM D7366-2008 基于回归方法从预测不确定度获得数据的标准指南

作者:标准资料网 时间:2024-05-11 03:46:34  浏览:9734   来源:标准资料网

【英文标准名称】:StandardPracticeforEstimationofMeasurementUncertaintyforDatafromRegression-basedMethods
【原文标准名称】:基于回归方法从预测不确定度获得数据的标准指南
【标准号】:ASTMD7366-2008
【标准状态】:现行
【国别】:美国
【发布日期】:2008
【实施或试行日期】:
【发布单位】:美国材料与试验协会(US-ASTM)
【起草单位】:D19.02
【标准类型】:(Practice)
【标准水平】:()
【中文主题词】:
【英文主题词】:measurementuncertainty;regression-basedmethods;calibration;predictioninterval;confidencelevel
【摘要】:Appropriateapplicationofthispracticeshouldresultinanestimateofthetest-methodx2019;suncertainty(atanyconcentrationwithintheworkingrange),whichcanbecomparedwithdata-qualityobjectivestoseeiftheuncertaintyisacceptable.Withdatasetsthatcomparerecoveredconcentrationwithtrueconcentration,theresultingregressionplotallowsthecorrectionoftherecoverydatatotruevalues.Reportingofsuchcorrectionsisatthediscretionoftheuser.Thispracticeshouldbeusedtoestimatethemeasurementuncertaintyforanyapplicationofatestmethodwheremeasurementuncertaintyisimportanttodatause.1.1ThispracticeestablishesastandardforcomputingthemeasurementuncertaintyforapplicabletestmethodsinCommitteeD19onWater.Thepracticedoesnotprovideasingle-pointestimatefortheentireworkingrange,butratherrelatestheuncertaintytoconcentration.Thestatisticaltechniqueofregressionisemployedduringdataanalysis.1.2Applicabletestmethodsarethosewhoseresultscomefromregression-basedmethodsandwhosedataareintra-laboratory(notinter-laboratorydata,suchasresultfromround-robinstudies).Foreachanalysisconductedusingsuchamethod,itisassumedthatafixed,reproducibleamountofsampleisintroduced.1.3Calculationofthemeasurementuncertaintyinvolvestheanalysisofdatacollectedtohelpcharacterizetheanalyticalmethodoveranappropriateconcentrationrange.Examplesourcesofdatainclude:1)calibrationstudies(whichmayormaynotbeconductedinpuresolvent),2)recoverystudies(whichtypicallyareconductedinmatrixandincludeallsample-preparationsteps),and3)collectionsofdataobtainedaspartofthemethodx2019;songoingQualityControlprogram.Useofmultipleinstruments,multipleoperators,orboth,andfield-samplingprotocolsmayormaynotbereflectedinthedata.1.4Inanydesignedstudywhosedataaretobeusedtocalculatemethoduncertainty,theusershouldthinkcarefullyaboutwhatthestudyistryingtoaccomplishandmuchvariationshouldbeincorporatedintothestudy.Generalguidanceondesigningstudies(forexample,calibration,recovery)isgiveninAppendixA.Detailedguidelinesonsourcesofvariationareoutsidethescopeofthispractice,butgeneralpointstoconsiderareincludedinAppendixB,whichisnotintendedtobeexhaustive.Withanystudy,theusermustthinkcarefullyaboutthefactorsinvolvedwithconductingtheanalysis,andmustrealizethatthecomputedmeasurementuncertaintywillreflectthequalityoftheinputdata.1.5Associatedwiththemeasurementuncertaintyisauser-chosenlevelofstatisticalconfidence.1.6Atanyconcentrationintheworkingrange,themeasurementuncertaintyisplus-or-minusthehalf-widthofthepredictionintervalassociatedwiththeregressionline.1.7Itisassumedthattheuserhasaccesstoastatisticalsoftwarepackageforperformingregression.Astatisticianshouldbeconsultedifassistanceisneededinselectingsuchaprogram.1.8Astatisticianalsoshouldbeconsultedifdatatransformationsarebeingconsidered.1.9Thisstandarddoesnotpurporttoaddressallofthesafetyconcerns,ifany,associatedwithitsuse.Itistheresponsibilityoftheuserofthisstandardtoestablishappropriatesafetyandhealthpracticesanddeterminetheapplicabilityofregulatorylimitationspriortouse.
【中国标准分类号】:Z16
【国际标准分类号】:13_060_45
【页数】:7P.;A4
【正文语种】:英语