資源描述:
《Receiver Operating Characteristic.pdf》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、ReceiverOperatingCharacteristic(ROC)CurvePreparation-ATutorialMassachusettsWaterResourcesAuthorityEnvironmentalQualityDepartmentReportENQUAD2005-20CitationMorrison,AnnMichelle.2005.ReceiverOperatingCharacteristic(ROC)CurvePreparation-ATutorial.Boston:Massachus
2、ettsWaterResourcesAuthority.ReportENQUAD2005-20.5p.2005-20ROCCurvePreparationReceiverOperatingCharacteristic(ROC)CurvePreparation–AtutorialReceiverOperatingCharacteristic(ROC)curvesweredevelopedinthefieldofstatisticaldecisiontheory,andlaterusedinthefieldofsign
3、aldetectionforanalyzingradarimagesduringWorldWarII(1).ROCcurvesenabledradaroperatorstodistinguishbetweenanenemytarget,afriendlyship,ornoise.ROCcurvesassessthevalueofdiagnostictestsbyprovidingastandardmeasureoftheabilityofatesttocorrectlyclassifysubjects.Thebio
4、medicalfieldusesROCcurvesextensivelytoassesstheefficacyofdiagnostictestsindiscriminatingbetweenhealthyanddiseasedindividuals(2).ROCcurvescan(1)assesstheoveralldiscriminatoryabilityofdifferentpotentialindicatorvariablesbygeneratingacommonmetricforcomparisonand(
5、2)aidintheselectionofaspecificvalueofanindicatorvariabletouseasathreshold,orlimit,thatprovidesadesiredtrade-offinthetruepositiverateandfalsepositiverate.Withrespecttobeachwaterqualityindicatorvariables,ROCcurvescanquantifytheoveralleffectivenessofdifferentindi
6、catorvariablestocorrectlyandincorrectlyclassifyabeachassuitableforswimmingandgenerateasinglemetricbywhichthedifferentindicatorvariablescanbecompared.Statisticaltheorystatesthattherearetwopopulationsofwater:thatwhichissuitableforswimmingandthatwhichisunsuitable
7、forswimming.Thedistributionsofthesepopulationsoverlap.Whenyouchooseanindicator,suchasEnterococcusdensitytodefinethetransitionbetweenwaterthatissuitableforswimmingandwaterthatisunsuitableforswimming,tworegionsofinteresttoregulatorsareformed(Figure1).Thetrueposi
8、tiverate(TPR)correspondstotheproportionofwatersamplesunsuitableforswimmingthatarecorrectlyidentifiedbytheindicatorvariablethreshold,whichinFigure1isantecedentrainfall.Thefalseposit