ISYE 6501 MIDTERM 1 LATEST
2026/2027 ACTUAL EXAM QUESTIONS
WITH CORRECT ANSWERS 100%
VERIFIED GRADED A+
Rowsm-mCORRECTmANSWERmDatampointsmaremvaluesminmdatamtables
Columnsm-mCORRECTmANSWERmThem'answer'mformeachmdatampointm(response/outcome)
StructuredmDatam-mCORRECTmANSWERmQuantitative,mCategorical,mBinary,mUnrelated,mTimemSeries
UnstructuredmDatam-mCORRECTmANSWERmText
SupportmVectormModelm-
mCORRECTmANSWERmSupervisedmmachinemlearningmalgorithmmusedmformbothmclassificationmandmr
egressionmchallenges.m
Mostlymusedminmclassificationmproblemsmbymplottingmeachmdatamitemmasmampointminmn-
dimensionalmspacem(nmismthemnumbermofmfeaturesmyoumhave)mwithmthemvaluemofmeachmfeaturembei
ngmthemvaluemofmamparticularmcoordinate.m
Thenmyoumclassifymbymfindingmamhyperplanemthatmdifferentiatesmthem2mclassesmverymwell.mSupportm
vectorsmaremsimplymthemcoordinatesmofmindividualmobservationm--
mitmbestmsegregatesmthemtwomclassesm(hyperplanem/mline).
WhatmdomyoumwantmtomfindmwithmamSVMmmodel?m-
mCORRECTmANSWERmFindmvaluesmofma0,ma1,...,upmtomammthatmclassifiesmthempointsmcorrectlymandm
hasmthemmaximummgapmormmarginmbetweenmthemparallelmlines.
,WhatmshouldmthemsummofmthemgreenmpointsminmamSVMmmodelmbe?m-
mCORRECTmANSWERmThemsummofmgreenmpointsmshouldmbemgreatermthanmormequalmtom1
WhatmshouldmthemsummofmthemredmpointsminmamSVMmmodelmbe?m-
mCORRECTmANSWERmThemsummofmredmpointsmshouldmbemlessmthanmormequalmtom-1
Whatmshouldmthemtotalmsummofmgreenmandmredmpointsmbe?m-
mCORRECTmANSWERmThemtotalmsummofmallmgreenmandmredmpointsmshouldmbemequalmtomormgreaterm
thanm1mbecausemyjmism1mformgreenmandm-1mformred.
Firstmprincipalmcomponentm-mCORRECTmANSWERmPCAm--
mamlinearmcombinationmofmoriginalmpredictormvariablesmwhichmcapturesmthemmaximummvariancemin
mthemdatamset.mItmdeterminesmthemdirectionmofmhighestmvariabilityminmthemdata.mLargermthemvariabil
itymcapturedminmfirstmcomponent,mlargermtheminformationmcapturedmbymcomponent.mNomothermco
mponentmcanmhavemvariabilitymhighermthanmfirstmprincipalmcomponent.
itmminimizesmthemsummofmsquaredmdistancembetweenmamdatampointmandmthemline.
Secondmprincipalmcomponentm-mCORRECTmANSWERmPCAm--
malsomamlinearmcombinationmofmoriginalmpredictorsmwhichmcapturesmthemremainingmvarianceminmth
emdatamsetmandmismuncorrelatedmwithmZ¹.mInmothermwords,mthemcorrelationmbetweenmfirstmandmsec
ondmcomponentmshouldmismzero.
Whatmifmit'smnotmpossiblemtomseparatemgreenmandmredmpointsminmamSVMmmodel?m-
mCORRECTmANSWERmUtilizemamsoftmclassifierm--
mInmamsoftmclassificationmcontext,mwemmightmaddmanmextrammultipliermformeachmtypemofmerrormwithm
amlargermpenalty,mthemlessmwemwantmtomacceptmmis-classifyingmthatmtypemofmpoint.
SoftmClassifierm-
mCORRECTmANSWERmAccountmformerrorsminmSVMmclassification.mTradingmoffmminimizingmerrorsmwe
mmakemandmmaximizingmthemmargin.
, Tomtrademoffmbetweenmthem,mwempickmamlambdamvaluemandmminimizemamcombinationmofmerrorman
dmmargin.mAsmlambdamgetsmlarge,mthismtermmgetsmlarge.
Themimportancemofmamlargemmarginmoutweighsmavoidingmmistakesmandmclassifyingmknownmdatamp
oints.
ShouldmyoumscalemyourmdataminmamSVMmmodel?m-
mCORRECTmANSWERmYes,msomthemordersmofmmagnitudemaremapproximatelymthemsame.
Datammustmbeminmboundedmrange.
Commonmscaling:mdatambetweenm0mandm1
a.mScalemfactormbymfactor
b.mLinearly
HowmshouldmyoumfindmwhichmcoefficientsmtomholdmvalueminmamSVMmmodel?m-
mCORRECTmANSWERmIfmtheremismamcoefficientmwho'smvaluemismverymclosemtom0,mmeansmthemcorresp
ondingmattributemismprobablymnotmrelevantmformclassification.
DoesmSVMmworkmthemsamemformmultiplemdimensions?m-mCORRECTmANSWERmYes
DoesmamSVMmclassifiermneedmtombemamstraightmline?m-
mCORRECTmANSWERmNo,mSVMmcanmbemgeneralizedmusingmkernelmmethodsmthatmallowmformnonline
armclassifiers.mSoftwaremhasmamkernelmSVMmfunctionmthatmyoumcanmusemtomsolvemformbothmlinearma
ndmnonlinearmclassifiers.
CanmclassificationmquestionsmbemansweredmasmprobabilitiesminmSVM?m-mCORRECTmANSWERmYes.
KmNearestmNeighbormAlgorithmm-
mCORRECTmANSWERmFindmthemclassmofmthemnewmpoint,mPickmthemkmclosestmpointsmtomthemnewmon
e,mthemnewmpointsmclassmismthemmostmcommonmamongstmthemkmneighbors.
2026/2027 ACTUAL EXAM QUESTIONS
WITH CORRECT ANSWERS 100%
VERIFIED GRADED A+
Rowsm-mCORRECTmANSWERmDatampointsmaremvaluesminmdatamtables
Columnsm-mCORRECTmANSWERmThem'answer'mformeachmdatampointm(response/outcome)
StructuredmDatam-mCORRECTmANSWERmQuantitative,mCategorical,mBinary,mUnrelated,mTimemSeries
UnstructuredmDatam-mCORRECTmANSWERmText
SupportmVectormModelm-
mCORRECTmANSWERmSupervisedmmachinemlearningmalgorithmmusedmformbothmclassificationmandmr
egressionmchallenges.m
Mostlymusedminmclassificationmproblemsmbymplottingmeachmdatamitemmasmampointminmn-
dimensionalmspacem(nmismthemnumbermofmfeaturesmyoumhave)mwithmthemvaluemofmeachmfeaturembei
ngmthemvaluemofmamparticularmcoordinate.m
Thenmyoumclassifymbymfindingmamhyperplanemthatmdifferentiatesmthem2mclassesmverymwell.mSupportm
vectorsmaremsimplymthemcoordinatesmofmindividualmobservationm--
mitmbestmsegregatesmthemtwomclassesm(hyperplanem/mline).
WhatmdomyoumwantmtomfindmwithmamSVMmmodel?m-
mCORRECTmANSWERmFindmvaluesmofma0,ma1,...,upmtomammthatmclassifiesmthempointsmcorrectlymandm
hasmthemmaximummgapmormmarginmbetweenmthemparallelmlines.
,WhatmshouldmthemsummofmthemgreenmpointsminmamSVMmmodelmbe?m-
mCORRECTmANSWERmThemsummofmgreenmpointsmshouldmbemgreatermthanmormequalmtom1
WhatmshouldmthemsummofmthemredmpointsminmamSVMmmodelmbe?m-
mCORRECTmANSWERmThemsummofmredmpointsmshouldmbemlessmthanmormequalmtom-1
Whatmshouldmthemtotalmsummofmgreenmandmredmpointsmbe?m-
mCORRECTmANSWERmThemtotalmsummofmallmgreenmandmredmpointsmshouldmbemequalmtomormgreaterm
thanm1mbecausemyjmism1mformgreenmandm-1mformred.
Firstmprincipalmcomponentm-mCORRECTmANSWERmPCAm--
mamlinearmcombinationmofmoriginalmpredictormvariablesmwhichmcapturesmthemmaximummvariancemin
mthemdatamset.mItmdeterminesmthemdirectionmofmhighestmvariabilityminmthemdata.mLargermthemvariabil
itymcapturedminmfirstmcomponent,mlargermtheminformationmcapturedmbymcomponent.mNomothermco
mponentmcanmhavemvariabilitymhighermthanmfirstmprincipalmcomponent.
itmminimizesmthemsummofmsquaredmdistancembetweenmamdatampointmandmthemline.
Secondmprincipalmcomponentm-mCORRECTmANSWERmPCAm--
malsomamlinearmcombinationmofmoriginalmpredictorsmwhichmcapturesmthemremainingmvarianceminmth
emdatamsetmandmismuncorrelatedmwithmZ¹.mInmothermwords,mthemcorrelationmbetweenmfirstmandmsec
ondmcomponentmshouldmismzero.
Whatmifmit'smnotmpossiblemtomseparatemgreenmandmredmpointsminmamSVMmmodel?m-
mCORRECTmANSWERmUtilizemamsoftmclassifierm--
mInmamsoftmclassificationmcontext,mwemmightmaddmanmextrammultipliermformeachmtypemofmerrormwithm
amlargermpenalty,mthemlessmwemwantmtomacceptmmis-classifyingmthatmtypemofmpoint.
SoftmClassifierm-
mCORRECTmANSWERmAccountmformerrorsminmSVMmclassification.mTradingmoffmminimizingmerrorsmwe
mmakemandmmaximizingmthemmargin.
, Tomtrademoffmbetweenmthem,mwempickmamlambdamvaluemandmminimizemamcombinationmofmerrorman
dmmargin.mAsmlambdamgetsmlarge,mthismtermmgetsmlarge.
Themimportancemofmamlargemmarginmoutweighsmavoidingmmistakesmandmclassifyingmknownmdatamp
oints.
ShouldmyoumscalemyourmdataminmamSVMmmodel?m-
mCORRECTmANSWERmYes,msomthemordersmofmmagnitudemaremapproximatelymthemsame.
Datammustmbeminmboundedmrange.
Commonmscaling:mdatambetweenm0mandm1
a.mScalemfactormbymfactor
b.mLinearly
HowmshouldmyoumfindmwhichmcoefficientsmtomholdmvalueminmamSVMmmodel?m-
mCORRECTmANSWERmIfmtheremismamcoefficientmwho'smvaluemismverymclosemtom0,mmeansmthemcorresp
ondingmattributemismprobablymnotmrelevantmformclassification.
DoesmSVMmworkmthemsamemformmultiplemdimensions?m-mCORRECTmANSWERmYes
DoesmamSVMmclassifiermneedmtombemamstraightmline?m-
mCORRECTmANSWERmNo,mSVMmcanmbemgeneralizedmusingmkernelmmethodsmthatmallowmformnonline
armclassifiers.mSoftwaremhasmamkernelmSVMmfunctionmthatmyoumcanmusemtomsolvemformbothmlinearma
ndmnonlinearmclassifiers.
CanmclassificationmquestionsmbemansweredmasmprobabilitiesminmSVM?m-mCORRECTmANSWERmYes.
KmNearestmNeighbormAlgorithmm-
mCORRECTmANSWERmFindmthemclassmofmthemnewmpoint,mPickmthemkmclosestmpointsmtomthemnewmon
e,mthemnewmpointsmclassmismthemmostmcommonmamongstmthemkmneighbors.
