All Chapters Covered
M M
SOLUTION MANUAL
M
,Contents
1 Introduction 3
1.11M Exercises .............................................................................................................................................................. 3
2 DataM Preprocessing 13
2.8 Exercises ...............................................................................................................................................................13
3 DataM WarehouseM andM OLAPM Technology:M AnM Overview 31
3.7 Exercises ...............................................................................................................................................................31
4 DataM CubeM ComputationM andM DataM Generalization 41
4.5 Exercises ...............................................................................................................................................................41
5 MiningM FrequentM Patterns,M Associations,M andM Correlations 53
5.7 Exercises ...............................................................................................................................................................53
6 ClassificationMandM Prediction 69
6.17M Exercises .............................................................................................................................................................69
7 ClusterM Analysis 79
7.13M Exercises .............................................................................................................................................................79
8 MiningM Stream,M Time-Series,M andM SequenceM Data 91
8.6 Exercises ...............................................................................................................................................................91
9 GraphM Mining,M SocialM NetworkM Analysis,M andM MultirelationalM DataM Mining 103
9.5 Exercises .............................................................................................................................................................103
10 MiningM Object,M Spatial,M Multimedia,M Text,M andM WebM Data 111
10.7 Exercises .............................................................................................................................................................111
11 ApplicationsM andM TrendsM inM DataM Mining 123
11.7 Exercises .............................................................................................................................................................123
1
,Chapter 1 M
Introduction
1.11 Exercises
1.1. WhatMisMdataMminingM?M InMyourManswer,MaddressMtheMfollowing:
(a) IsM itM anotherM hype?
(b) IsM itM aM simpleM transformationM ofM technologyM developedM fromM databases,M statistics,M andM machineM learning
?
(c) ExplainM howM theM evolutionM ofM databaseM technologyM ledM toM dataM mining.
(d) DescribeM theM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discovery.
Answer:
DataMminingMrefersMtoMtheMprocessMorMmethodMthatMextractsMorM“mines”MinterestingMknowledge
MorMpatternsMfromMlargeMamountsMofMdata.
(a) IsM itM anotherM hype?
DataMminingMisMnotManotherMhype.M Instead,M theMneedMforMdataMminingMhasMarisenMdueMtoMtheM
wideM availabilityMofMhugeMamountsMofMdataMandMtheMimminentMneedMforMturningMsuchMdataMinto
MusefulMinformationMandMknowledge.M Thus,MdataMminingMcanMbeMviewedMasMtheMresultMofMtheMn
aturalMevolutionMofMinformationMtechnology.
(b) IsMitMaMsimpleMtransformationMofMtechnologyMdevelopedMfromMdatabases,Mstatistics,MandMmachin
eMlearning?MNo.M DataMminingMisMmoreMthanMaMsimpleMtransformationMofMtechnologyMdevelop
edMfromMdatabases,Msta-
Mtistics,M andM machineM learning. M Instead,M dataM miningM involvesM anM integration, M ratherM th
anM aM simple
transformation,M ofM techniquesM fromM multipleM disciplinesM suchM asM databaseM technology,M statistics,M ma
-
chineMlearning,Mhigh-
performanceMcomputing,MpatternMrecognition,MneuralMnetworks,MdataMvisualization,MinformationM ret
rieval,M imageM andM signalM processing,M andM spatialM dataM analysis.
(c) ExplainM howM theM evolutionM ofM databaseM technologyM ledM toM dataM mining.
DatabaseMtechnologyMbeganMwithMtheMdevelopmentMofMdataMcollectionMandMdatabaseMcreationM
mechanismsMthatMledMtoMtheMdevelopment MofMeffectiveMmechanismsMforMdataMmanagementMin
cludingMdataMstorageMandMretrieval,MandMqueryMandMtransactionMprocessing.MTheMlargeMnumbe
rMofMdatabaseMsystemsMofferingMqueryMandMtransactionMprocessingMeventuallyMandMnaturallyMle
dMtoMtheMneedMforMdataManalysisMandMunderstanding.MHence,MdataMminingMbeganMitsMdevelop
mentMoutMofMthisMnecessity.
(d) DescribeM theM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discovery.
TheM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discoveryM areM asM follo
ws:
• DataMcleaning,MaMprocessMthatMremovesMorMtransformsMnoiseMandMinconsistentMdata
• DataM integration,M whereM multipleM dataM sourcesM mayM beM combined
3
, 4 CHAPTERM 1.M M INTRODUCTI
ON
• DataMselection,MwhereMdataMrelevantMtoMtheManalysisMtaskMareMretrievedMfromMtheMdatabase
• DataM transformation,M whereM dataM areM transformedM orM consolidatedM intoM formsM appro
priateM forMmining
• DataMmining,ManMessentialMprocessMwhereMintelligentMandMefficientMmethodsMareMapplied
MinMorderMtoMextractMpatterns
• PatternM evaluation,M aM processM thatM identifiesM theM trulyM interestingM patternsM representin
gM knowl-MedgeMbasedMonMsomeMinterestingnessMmeasures
• KnowledgeM presentation,M whereM visualizationM andM knowledgeM representationM techniquesM ar
eM usedMtoMpresentMtheMminedMknowledgeMtoMtheMuser
1.2. PresentManMexampleMwhere MdataMminingMisMcrucialMtoMtheMsuccessMofMaMbusiness.M WhatMdataMmini
ngMfunctionsMdoesMthisMbusinessMneed?M CanMtheyMbeMperformedMalternativelyMbyMdataMqueryMproce
ssingMorMsimpleMstatisticalManalysis?
Answer:
AM departmentM store,M forM example,M canM useM dataM miningM toM assistM withM itsM targetM marketingM m
ailM campaign.MUsingMdataMminingMfunctionsMsuchMasMassociation,MtheMstoreMcanMuseMtheMminedMstron
gMassociationMrulesMtoMdetermineM whichM productsM boughtM byM oneM groupM ofM customersM areM likely
M toM leadM toM theM buying M ofM certainMotherMproducts.M WithMthisMinformation, MtheMstoreMcanMthenMm
ailMmarketingMmaterialsMonlyMtoMthoseMkindsMofMcustomers M whoM exhibitM aM highM likelihoodM ofM purc
hasingM additionalM products.M DataM queryM processingM isM usedMforMdataMorMinformationMretrievalMandM
doesMnotMhaveMtheMmeansMforMfindingMassociationMrules.M Similarly,MsimpleMstatisticalManalysisMcannot
MhandleMlargeMamountsMofMdataMsuchMasMthoseMofMcustomerMrecordsMinMaMdepartment M store.
1.3. SupposeMyourMtaskMasMaMsoftwareMengineerMatMBig-
UniversityMisMtoMdesignMaMdataMminingMsystemMtoMexamineMtheirMuniversityMcourseMdatabase,Mw
hichMcontainsMtheMfollowingMinformation:M theMname,Maddress,MandMstatusM(e.g.,MundergraduateMor
Mgraduate)Mof MeachMstudent,MtheMcoursesMtaken,MandMtheirMcumulativeMgradeMpoint MaverageM(GPA)
.MDescribeMtheMarchitectureMyouMwouldMchoose.M WhatMisMtheMpurposeMofMeachMcomponentMofMthisMarch
itecture?
Answer:
AM dataM miningM architectureM thatM canM beM usedM forM thisM applicationM wouldM consistM ofM theM follo
wingM majorM components:
• AMdatabase,MdataMwarehouse,MorMotherMinformationMrepository,MwhichMconsistsMofMthe
MsetMofMdatabases,MdataMwarehouses,Mspreadsheets,MorMotherMkinds MofMinformation MrepositoriesMco
ntainingMtheMstudentMandMcourseMinformation.
• AMdatabaseMorMdataMwarehouseMserver,MwhichMfetchesMtheMrelevantMdataMbasedMonMtheMu
sers’MdataMminingMrequests.
• AMknowledgeMbaseMthatMcontainsMtheMdomainMknowledgeMusedMtoMguideMtheMsearchMorMtoMev
aluateMtheMinterestingnessMofMresultingMpatterns.M ForMexample,MtheMknowledgeMbaseMmayMcontain
MconceptMhierarchies MandM metadataM (e.g.,M describing M dataM fromM multipleM heterogeneous M sourc
es).
• AMdataMminingMengine,MwhichMconsistsMofMaMsetMofMfunctionalMmodulesMforMtasksMsuchMasM
classification,Massociation,M classification,M clusterM analysis,M andM evolutionM andM deviationM analysi
s.
• AMpatternMevaluationMmoduleMthatMworksMinMtandemMwithMtheMdataMminingMmodulesMbyMe
mployingMinterestingnessM measuresM toM helpM focusM theM searchM towardsM interestingM patterns.
• AMgraphicalMuserMinterfaceMthatMprovidesMtheMuserMwithManMinteractive MapproachMtoMthe
MdataMminingMsystem.
M M
SOLUTION MANUAL
M
,Contents
1 Introduction 3
1.11M Exercises .............................................................................................................................................................. 3
2 DataM Preprocessing 13
2.8 Exercises ...............................................................................................................................................................13
3 DataM WarehouseM andM OLAPM Technology:M AnM Overview 31
3.7 Exercises ...............................................................................................................................................................31
4 DataM CubeM ComputationM andM DataM Generalization 41
4.5 Exercises ...............................................................................................................................................................41
5 MiningM FrequentM Patterns,M Associations,M andM Correlations 53
5.7 Exercises ...............................................................................................................................................................53
6 ClassificationMandM Prediction 69
6.17M Exercises .............................................................................................................................................................69
7 ClusterM Analysis 79
7.13M Exercises .............................................................................................................................................................79
8 MiningM Stream,M Time-Series,M andM SequenceM Data 91
8.6 Exercises ...............................................................................................................................................................91
9 GraphM Mining,M SocialM NetworkM Analysis,M andM MultirelationalM DataM Mining 103
9.5 Exercises .............................................................................................................................................................103
10 MiningM Object,M Spatial,M Multimedia,M Text,M andM WebM Data 111
10.7 Exercises .............................................................................................................................................................111
11 ApplicationsM andM TrendsM inM DataM Mining 123
11.7 Exercises .............................................................................................................................................................123
1
,Chapter 1 M
Introduction
1.11 Exercises
1.1. WhatMisMdataMminingM?M InMyourManswer,MaddressMtheMfollowing:
(a) IsM itM anotherM hype?
(b) IsM itM aM simpleM transformationM ofM technologyM developedM fromM databases,M statistics,M andM machineM learning
?
(c) ExplainM howM theM evolutionM ofM databaseM technologyM ledM toM dataM mining.
(d) DescribeM theM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discovery.
Answer:
DataMminingMrefersMtoMtheMprocessMorMmethodMthatMextractsMorM“mines”MinterestingMknowledge
MorMpatternsMfromMlargeMamountsMofMdata.
(a) IsM itM anotherM hype?
DataMminingMisMnotManotherMhype.M Instead,M theMneedMforMdataMminingMhasMarisenMdueMtoMtheM
wideM availabilityMofMhugeMamountsMofMdataMandMtheMimminentMneedMforMturningMsuchMdataMinto
MusefulMinformationMandMknowledge.M Thus,MdataMminingMcanMbeMviewedMasMtheMresultMofMtheMn
aturalMevolutionMofMinformationMtechnology.
(b) IsMitMaMsimpleMtransformationMofMtechnologyMdevelopedMfromMdatabases,Mstatistics,MandMmachin
eMlearning?MNo.M DataMminingMisMmoreMthanMaMsimpleMtransformationMofMtechnologyMdevelop
edMfromMdatabases,Msta-
Mtistics,M andM machineM learning. M Instead,M dataM miningM involvesM anM integration, M ratherM th
anM aM simple
transformation,M ofM techniquesM fromM multipleM disciplinesM suchM asM databaseM technology,M statistics,M ma
-
chineMlearning,Mhigh-
performanceMcomputing,MpatternMrecognition,MneuralMnetworks,MdataMvisualization,MinformationM ret
rieval,M imageM andM signalM processing,M andM spatialM dataM analysis.
(c) ExplainM howM theM evolutionM ofM databaseM technologyM ledM toM dataM mining.
DatabaseMtechnologyMbeganMwithMtheMdevelopmentMofMdataMcollectionMandMdatabaseMcreationM
mechanismsMthatMledMtoMtheMdevelopment MofMeffectiveMmechanismsMforMdataMmanagementMin
cludingMdataMstorageMandMretrieval,MandMqueryMandMtransactionMprocessing.MTheMlargeMnumbe
rMofMdatabaseMsystemsMofferingMqueryMandMtransactionMprocessingMeventuallyMandMnaturallyMle
dMtoMtheMneedMforMdataManalysisMandMunderstanding.MHence,MdataMminingMbeganMitsMdevelop
mentMoutMofMthisMnecessity.
(d) DescribeM theM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discovery.
TheM stepsM involvedM inM dataM miningM whenM viewedM asM aM processM ofM knowledgeM discoveryM areM asM follo
ws:
• DataMcleaning,MaMprocessMthatMremovesMorMtransformsMnoiseMandMinconsistentMdata
• DataM integration,M whereM multipleM dataM sourcesM mayM beM combined
3
, 4 CHAPTERM 1.M M INTRODUCTI
ON
• DataMselection,MwhereMdataMrelevantMtoMtheManalysisMtaskMareMretrievedMfromMtheMdatabase
• DataM transformation,M whereM dataM areM transformedM orM consolidatedM intoM formsM appro
priateM forMmining
• DataMmining,ManMessentialMprocessMwhereMintelligentMandMefficientMmethodsMareMapplied
MinMorderMtoMextractMpatterns
• PatternM evaluation,M aM processM thatM identifiesM theM trulyM interestingM patternsM representin
gM knowl-MedgeMbasedMonMsomeMinterestingnessMmeasures
• KnowledgeM presentation,M whereM visualizationM andM knowledgeM representationM techniquesM ar
eM usedMtoMpresentMtheMminedMknowledgeMtoMtheMuser
1.2. PresentManMexampleMwhere MdataMminingMisMcrucialMtoMtheMsuccessMofMaMbusiness.M WhatMdataMmini
ngMfunctionsMdoesMthisMbusinessMneed?M CanMtheyMbeMperformedMalternativelyMbyMdataMqueryMproce
ssingMorMsimpleMstatisticalManalysis?
Answer:
AM departmentM store,M forM example,M canM useM dataM miningM toM assistM withM itsM targetM marketingM m
ailM campaign.MUsingMdataMminingMfunctionsMsuchMasMassociation,MtheMstoreMcanMuseMtheMminedMstron
gMassociationMrulesMtoMdetermineM whichM productsM boughtM byM oneM groupM ofM customersM areM likely
M toM leadM toM theM buying M ofM certainMotherMproducts.M WithMthisMinformation, MtheMstoreMcanMthenMm
ailMmarketingMmaterialsMonlyMtoMthoseMkindsMofMcustomers M whoM exhibitM aM highM likelihoodM ofM purc
hasingM additionalM products.M DataM queryM processingM isM usedMforMdataMorMinformationMretrievalMandM
doesMnotMhaveMtheMmeansMforMfindingMassociationMrules.M Similarly,MsimpleMstatisticalManalysisMcannot
MhandleMlargeMamountsMofMdataMsuchMasMthoseMofMcustomerMrecordsMinMaMdepartment M store.
1.3. SupposeMyourMtaskMasMaMsoftwareMengineerMatMBig-
UniversityMisMtoMdesignMaMdataMminingMsystemMtoMexamineMtheirMuniversityMcourseMdatabase,Mw
hichMcontainsMtheMfollowingMinformation:M theMname,Maddress,MandMstatusM(e.g.,MundergraduateMor
Mgraduate)Mof MeachMstudent,MtheMcoursesMtaken,MandMtheirMcumulativeMgradeMpoint MaverageM(GPA)
.MDescribeMtheMarchitectureMyouMwouldMchoose.M WhatMisMtheMpurposeMofMeachMcomponentMofMthisMarch
itecture?
Answer:
AM dataM miningM architectureM thatM canM beM usedM forM thisM applicationM wouldM consistM ofM theM follo
wingM majorM components:
• AMdatabase,MdataMwarehouse,MorMotherMinformationMrepository,MwhichMconsistsMofMthe
MsetMofMdatabases,MdataMwarehouses,Mspreadsheets,MorMotherMkinds MofMinformation MrepositoriesMco
ntainingMtheMstudentMandMcourseMinformation.
• AMdatabaseMorMdataMwarehouseMserver,MwhichMfetchesMtheMrelevantMdataMbasedMonMtheMu
sers’MdataMminingMrequests.
• AMknowledgeMbaseMthatMcontainsMtheMdomainMknowledgeMusedMtoMguideMtheMsearchMorMtoMev
aluateMtheMinterestingnessMofMresultingMpatterns.M ForMexample,MtheMknowledgeMbaseMmayMcontain
MconceptMhierarchies MandM metadataM (e.g.,M describing M dataM fromM multipleM heterogeneous M sourc
es).
• AMdataMminingMengine,MwhichMconsistsMofMaMsetMofMfunctionalMmodulesMforMtasksMsuchMasM
classification,Massociation,M classification,M clusterM analysis,M andM evolutionM andM deviationM analysi
s.
• AMpatternMevaluationMmoduleMthatMworksMinMtandemMwithMtheMdataMminingMmodulesMbyMe
mployingMinterestingnessM measuresM toM helpM focusM theM searchM towardsM interestingM patterns.
• AMgraphicalMuserMinterfaceMthatMprovidesMtheMuserMwithManMinteractive MapproachMtoMthe
MdataMminingMsystem.
