ISYE 6414 FINAL EXAM REVIEW
QUESTIONS AND VERIFIED ANSWERS
2024
LeastdSquaredEliminationd(LSE)dcannotdbedapplieddtodGLMdmodels.d-dans✔✔Falsed-
ditdisdapplicabledbutddoesdnotduseddataddistributiondinformationdfully.
Indmultipledlineardregressiondwithdidddanddequaldvariance,dthedleastdsquaresdestimationdof
dregressiondcoefficientsdaredalwaysdunbiased.d-dans✔✔Trued-
dthedleastdsquaresdestimatesdaredBLUEd(BestdLineardUnbiaseddEstimates)dindmultipledline
ardregression.
MaximumdLikelihooddEstimationdisdnotdapplicabledfordsimpledlineardregressiondanddmultipl
edlineardregression.d-dans✔✔Falsed-
dIndSLRdanddMLR,dthedSLEdanddMLEdaredthedsamedwithdnormaldiddddata.
Thedbackwarddeliminationdrequiresdadpre-setdprobabilitydofdtypedIIderrord-dans✔✔Falsed-
dTypedIderror
ThedfirstddegreedofdfreedomdindthedFddistributiondfordanydofdthedthreedproceduresdindstepwi
sedisdalwaysdequaldtodone.d-dans✔✔True
MLEdisduseddfordthedGLMsdfordhandlingdcomplicateddlinkdfunctiondmodelingdindthedX-
Ydrelationship.d-dans✔✔True
IndthedGLMsdthedlinkdfunctiondcannotdbedadnondlineardregression.d-dans✔✔Falsed-
dItdcandbedlinear,dnondlinear,dordparametric
, Whendthedp-valuedofdthedslopedestimatedindthedSLRdisdsmalldthedr-
squareddbecomesdsmallerdtoo.d-dans✔✔Falsed-
dWhendPdvaluedisdsmall,dthedmodeldfitsdbecomedmoredsignificantdanddRdsquareddbecomedl
arger.
IndGLMsdthedmaindreasondoneddoesdnotdusedLSEdtodestimatedmodeldparametersdisdthedpo
tentialdconstraineddindthedparameters.d-dans✔✔Falsed-
dThedpotentialdconstraintdindthedparametersdofdGLMsdisdhandleddbydthedlinkdfunction.
ThedR-squareddanddadjusteddR-
squareddarednotdappropriatedmodeldcomparisonsdfordnondlineardregressiondbutdaredfordlin
eardregressiondmodels.d-dans✔✔TRUEd-dThedunderlyingdassumptiondofdR-
squareddcalculationsdisdthatdyoudaredfittingdadlineardmodel.
TheddecisiondindusingdANOVAdtabledfordtestingdwhetherdadmodeldisdsignificantddependsdo
ndthednormalddistributiondofdthedresponsedvariabled-dans✔✔True
Whendtheddatadmaydnotdbednormallyddistributed,dAICdisdmoredappropriatedfordvariabledsel
ectiondthandadjusteddR-squaredd-dans✔✔True
Thedslopedofdadlineardregressiondequationdisdandexampledofdadcorrelationdcoefficient.d-
dans✔✔Falsed-dthedcorrelationdcoefficientdisdthedrdvalue.dWilldhavedthedsamed+dord-
dsigndasdthedslope.
Indmultipledlineardregression,dasdthedvaluedofdR-squareddincreases,dthedrelationship
betweendpredictorsdbecomesdstrongerd-dans✔✔Falsed-
drdsquareddmeasuresdhowdmuchdvariabilitydisdexplaineddbydthedmodel,dNOTdhowdstrongdth
edpredictorsdare.
Whenddealingdwithdadmultipledlineardregressiondmodel,dandadjusteddR-squareddcan
bedgreaterdthandthedcorrespondingdunadjusteddR-Squareddvalue.d-dans✔✔Falsed-
dthedadjusteddrsquareddvaluedtakedthednumberdanddtypesdofdpredictorsdintodaccount.dItdisdl
owerdthandthedrdsquareddvalue.
Indadmultipledregressiondproblem,dadquantitativedinputdvariabledxdisdreplaceddbydxd−
mean(x).dThedR-squareddfordthedfitteddmodeldwilldbedthedsamed-dans✔✔True
Thedestimateddcoefficientsdofdadregressiondlinedisdpositive,dwhendthedcoefficientdof
determinationdisdpositive.d-dans✔✔Falsed-drdsquareddisdalwaysdpositive.
Ifdthedoutcomedvariabledisdquantitativedanddalldexplanatorydvariablesdtakedvaluesd0dor
1,dadlogisticdregressiondmodeldisdmostdappropriate.d-dans✔✔Falsed-
dMoredresearchdisdnecessarydtoddeterminedthedcorrectdmodel.
Afterdfittingdadlogisticdregressiondmodel,dadplotdofdresidualsdversusdfitteddvaluesdis
QUESTIONS AND VERIFIED ANSWERS
2024
LeastdSquaredEliminationd(LSE)dcannotdbedapplieddtodGLMdmodels.d-dans✔✔Falsed-
ditdisdapplicabledbutddoesdnotduseddataddistributiondinformationdfully.
Indmultipledlineardregressiondwithdidddanddequaldvariance,dthedleastdsquaresdestimationdof
dregressiondcoefficientsdaredalwaysdunbiased.d-dans✔✔Trued-
dthedleastdsquaresdestimatesdaredBLUEd(BestdLineardUnbiaseddEstimates)dindmultipledline
ardregression.
MaximumdLikelihooddEstimationdisdnotdapplicabledfordsimpledlineardregressiondanddmultipl
edlineardregression.d-dans✔✔Falsed-
dIndSLRdanddMLR,dthedSLEdanddMLEdaredthedsamedwithdnormaldiddddata.
Thedbackwarddeliminationdrequiresdadpre-setdprobabilitydofdtypedIIderrord-dans✔✔Falsed-
dTypedIderror
ThedfirstddegreedofdfreedomdindthedFddistributiondfordanydofdthedthreedproceduresdindstepwi
sedisdalwaysdequaldtodone.d-dans✔✔True
MLEdisduseddfordthedGLMsdfordhandlingdcomplicateddlinkdfunctiondmodelingdindthedX-
Ydrelationship.d-dans✔✔True
IndthedGLMsdthedlinkdfunctiondcannotdbedadnondlineardregression.d-dans✔✔Falsed-
dItdcandbedlinear,dnondlinear,dordparametric
, Whendthedp-valuedofdthedslopedestimatedindthedSLRdisdsmalldthedr-
squareddbecomesdsmallerdtoo.d-dans✔✔Falsed-
dWhendPdvaluedisdsmall,dthedmodeldfitsdbecomedmoredsignificantdanddRdsquareddbecomedl
arger.
IndGLMsdthedmaindreasondoneddoesdnotdusedLSEdtodestimatedmodeldparametersdisdthedpo
tentialdconstraineddindthedparameters.d-dans✔✔Falsed-
dThedpotentialdconstraintdindthedparametersdofdGLMsdisdhandleddbydthedlinkdfunction.
ThedR-squareddanddadjusteddR-
squareddarednotdappropriatedmodeldcomparisonsdfordnondlineardregressiondbutdaredfordlin
eardregressiondmodels.d-dans✔✔TRUEd-dThedunderlyingdassumptiondofdR-
squareddcalculationsdisdthatdyoudaredfittingdadlineardmodel.
TheddecisiondindusingdANOVAdtabledfordtestingdwhetherdadmodeldisdsignificantddependsdo
ndthednormalddistributiondofdthedresponsedvariabled-dans✔✔True
Whendtheddatadmaydnotdbednormallyddistributed,dAICdisdmoredappropriatedfordvariabledsel
ectiondthandadjusteddR-squaredd-dans✔✔True
Thedslopedofdadlineardregressiondequationdisdandexampledofdadcorrelationdcoefficient.d-
dans✔✔Falsed-dthedcorrelationdcoefficientdisdthedrdvalue.dWilldhavedthedsamed+dord-
dsigndasdthedslope.
Indmultipledlineardregression,dasdthedvaluedofdR-squareddincreases,dthedrelationship
betweendpredictorsdbecomesdstrongerd-dans✔✔Falsed-
drdsquareddmeasuresdhowdmuchdvariabilitydisdexplaineddbydthedmodel,dNOTdhowdstrongdth
edpredictorsdare.
Whenddealingdwithdadmultipledlineardregressiondmodel,dandadjusteddR-squareddcan
bedgreaterdthandthedcorrespondingdunadjusteddR-Squareddvalue.d-dans✔✔Falsed-
dthedadjusteddrsquareddvaluedtakedthednumberdanddtypesdofdpredictorsdintodaccount.dItdisdl
owerdthandthedrdsquareddvalue.
Indadmultipledregressiondproblem,dadquantitativedinputdvariabledxdisdreplaceddbydxd−
mean(x).dThedR-squareddfordthedfitteddmodeldwilldbedthedsamed-dans✔✔True
Thedestimateddcoefficientsdofdadregressiondlinedisdpositive,dwhendthedcoefficientdof
determinationdisdpositive.d-dans✔✔Falsed-drdsquareddisdalwaysdpositive.
Ifdthedoutcomedvariabledisdquantitativedanddalldexplanatorydvariablesdtakedvaluesd0dor
1,dadlogisticdregressiondmodeldisdmostdappropriate.d-dans✔✔Falsed-
dMoredresearchdisdnecessarydtoddeterminedthedcorrectdmodel.
Afterdfittingdadlogisticdregressiondmodel,dadplotdofdresidualsdversusdfitteddvaluesdis
