数据仓库、数据发掘和联机分析处理.（影印版）

　　The last few years have seen a growing recognition of information as a key business tool. Those who successfully gather， analyze， understand， and act upon information are among the winners in this new information age. There- fore， it is only reasonable to expect the rate of producing and consuming infor- mation to grow. We can define information as that which resolves uncertainty. We can further say that decisionmaking is the progressive resolution of uncer- tainty and is a key to a purposeful behavior by any mechanism （or organism）. In general， the current business market dynamics make it abundantly clear that， for any company， information is the very key to survival.If we look at the evolution of the information processing technologies， we can see that while the first generation of client/server systems brought data to the desktop， not all of this data was easy to understand， unfortunately， and as such， it was not very useful to end users. As a result， a number of new tech- nologies have emerged that are focused on improving the information content of the data to empower the knowledge workers of today and tomorrow. Among these technologies are data warehousing， metadata repositories， online analyt- ical processing （OLAP）， and data mining. In some ways， these technologies are the manifestation of the maturity of the client/server computing model and its applicability to a wide variety of business problems. Therefore， this book is about the need， the value， and the technological means of acquiring and using information in the information age.

暂缺《数据仓库、数据发掘和联机分析处理.（影印版）》作者简介

Forewordxix
Prefacexxi
Part1.Foundation
Chapter1.IntroductiontoDataWarehousing
1.1WhyAlltheExcitement?
1.2TheNeedforDataWarehousing
1.3ParadigmShift
1.3.1ComputingParadigm
1.3.2BusinessParadigm
1.4BusinessProblemDefinition
1.5OperationalandInformationalDataStores*
1.6DataWarehouseDefinitionandCharacteristics
1.7DataWarehouseArchitecture
1.8ChapterSummary
Chapter2.Client/ServerComputingModelandDataWarehousing
2.1OverviewofClient/ServerArchitecture
2.1.1Host-BasedProcessing
2.1.2Master-SlaveProcessing
2.1.3First-GenerationClient/ServerProcessing
2.1.4Second-GenerationClient/ServerProcessing
2.2ServerSpecializationinClient/ServerComputingEnvironments
2.3ServerFunctions
2.4ServerHardwareArchitecture
2.5SystemConsiderations
2.6RISCversusClSC
2.7MultiprocessorSystems
2.7.1SMPDesign
2.7.2SMPFeatures
2.7.3SMPOperatingSystems
2.8SMPImplementations
Chapter3.ParallelProcessorsandClusterSystems
3.1Distributed-MemoryArchitecture
3.1.1Shared-NothingArchitectures
3.1.2Shared-DiskSystems
3.2ResearchIssues
3.3ClusterSystems
3.4AdvancesinMultiprocessingArchitectures
3.5OptimalHardwareArchitectureforQueryScalability*
3.5.1UniformityofDataAccessTimes
3.5.2SystemArchitectureTaxonomyandQueryExecution
3.6ServerOperatingSystems
3.6.1OperatingSystemRequirements
3.6.2MicrokernelTechnology
3.7OperatingSystemImplementations
3.7.1UNIX
3.7.2Windows/NT
3.7.3OS/2
3.7.4NetWare
3.7.5OSSummary
Chapter4.DistributedDBMSImplementations
4.1ImplementationTrendsandFeaturesofDistributedClient/ServerDBMS
4.1.1RDBMSArchitectureforScalability
4.1.2RDBMSPerformanceandEfficiencyFeatures
4.1.3TypesofParallelism
4.2DBMSConnectivity
4.3AdvancedRDBMSFeatures
4.4RDBMSReliabilityandAvailability
4.4.1Robustness,TransactionsRecovery,andConsistency
4.4.2FaultTolerance
4.5RDBMSAdministration
Chapter5.Client/ServerRDBMSSolutions
5.1State-of-the-MarketOverview
5.2Oracle
5.2.1SystemManagement
5.2.2OracleUniversalServer
5.2.3OracleConTextOption
5.2.4OracleSpatialDataOption
5.3Informix
5.3.1Features
5.3.2InformixUniversalServer
5.4Sybase
5.4.1SYBASESQLServer
5.4.2PerformanceImprovementsinSYBASESystem11
5.5IBM
5.5.1Background
5.5.2DB2UniversalDatabase
5.6Microsoft
5.6.1Background
5.6.2MSSQLServer
5.6.3DataWarehousingandMarketPositioning
Part2DataWarehousing
Chapter6.DataWarehousingComponents
6.1OverallArchitecture
6.2DataWarehouseDatabase
6.3Sourcing,Acquisition,Cleanup,andTransformationTools
6.4Metadata
6.5AccessTools
6.5.1QueryandReportingTools
6.5.2Applications
6.5.3OLAP
6.5.4DataMining
6.5.5DataVisualization
6.6DataMarts
6.7DataWarehouseAdministrationandManagement
6.8InformationDeliverySystem
Chapter7.BuildingaDatawarehouse
7.1BusinessConsiderations:ReturnonInvestment
7.1.1Approach
7.1.2OrganizationalIssues
7.2DesignConsiderations
7.2.1DataContent
7.2.2Metadata
7.2.3DataDistribution
7.2.4Tools
7.2.5PerformanceConsiderations
7.2.6NineDecisionsintheDesignofaDataWarehouse
7.3TechnicalConsiderations
7.3.1HardwarePlatforms
7.3.2DataWarehouseandDBMSSpecialization
7.3.3CommunicationsInfrastructure
7.4ImplementationConsiderations
7.4.1AccessTools
7.4.2DataExtraction,Cleanup,Transformation,andMigration
7.4.3DataPlacementStrategies
7.4.4Metadata
7.4.5UserSophisticationLevels
7.5IntegratedSolutions
7.6BenefitsofDataWarehousing
7.6.1TangibleBenefits
7.6.2IntangibleBenefits
Chapter8.MappingtheDataWarehousetoaMultiprocessorArchitecture
8.1RelationalDatabaseTechnologyforDataWarehouse
8.1.1TypesofParallelism
8.1.2DataPartitioning
8.2DatabaseArchitecturesforParallelProcessing
8.2.1Shared-MemoryArchitecture
8.2.2Shared-DiskArchitecture
8.2.3Shared-NothingArchitecture
8.2.4CombinedArchitecture
8.3ParallelRDBMSFeatures
8.4AlternativeTechnologies
8.5ParallelDBMSVendors
8.5.1Oracle
8.5.2Informix
8.5.3IBM
8.5.4Sybase
8.5.5Microsoft
8.5.6OtherRDBMSProducts
8.5.7SpecializedDatabaseProducts
Chapter9.DBMSSchemasforDecisionSupport
9.1DataLayoutforBestAccess
9.2MultidimensionalDataModel
9.3StarSchema
9.3.1DBAViewpoint
9.3.2PotentialPerformanceProblemswithStarSchemas
9.3.3SolutionstoPerformanceProblems
9.4STARjoinandSTARindex
9.5BitmappedIndexing
9.5.1SYBASEIQ
9.5.2Conclusion
9.6ColumnLocalStorage
9.7ComplexDataTypes
Chapter10.DataExtraction,Cleanup,andTransformationTools
10.1ToolRequirements
10.2VendorApproaches
10.3AccesstoLegacyData
10.4VendorSolutions
10.4.1PrismSolutions
10.4.2SASInstitute
10.4.3CarletonCorporation'sPassportandMetaCenter
10.4.4ValityCorporation
10.4.5EvolutionaryTechnologies
10.4.6InformationBuilders
10.5TransformationEngines
10.5.1Informatica
10.5.2Constellar
Chapter11.Metadata
11.1MetadataDefined
11.2MetadataInterchangeInitiative
11.3MetadataRepository
11.4MetadataManagement
11.5ImplementationExamples
11.5.1PlatinumRepository
11.5.2R&O:TheROCHADERepository
11.5.3PrismSolutions
11.5.4LogicWorksUniversalDirectory
11.6MetadataTrends
Part3.BusinessAnalysis
Chapter12.ReportingandQueryToolsandApplications
12.1ToolCategories
12.1.1ReportingTools
12.1.2ManagedQueryTools
12.1.3ExecutiveInformationSystemTools
12.1.4OLAPTools
12.1.5DataMiningTools
12.2TheNeedforApplications
12.3CognosImpromptu
12.4Applications
12.4.1PowerBuilder
12.4.2Forte
12.4.3InformationBuilders
Chapter13.On-LineAnalyticalProcessing(OLAP)
13.1NeedforOLAP
13.2MultidimensionalDataModel
13.3OLAPGuidelines
13.4MultidimensionalversusMulfirelationalOLAP
13.5CategorizationofOLAPTools
13.5.1MOLAP
13.5.2ROLAP
13.5.3ManagedQueryEnvironment(MQE)
13.6StateoftheMarket
13.6.1CognosPowerPlay
13.6.2IBIFOCUSFusion
13.6.3PilotSoftware
13.7OLAPToolsandtheInternet
13.8Conclusion
Chapter14.PatternsandModels
14.1Definitions
14.1.1WhatIsaPattern?WhatIsaModel?
14.1.2VisualizingaPattern
14.2ANoteonTerminology
14.3WhereAreModelsUsed?
14.3.1Problem1:Selection
14.3.2Problem2:Acquisition
14.3.3Problem3:Retention
14.3.4Problem4:Extension
14.4WhatIsthe"Right"Model?
14.4.1ThePerfectModel
14.4.2MissingData
14.5Sampling
14.5.1TheNecessityofSampling
14.5.2RandomSampling
14.6ExperimentalDesign
14.6.1AvoidingBias
14.6.2MoreonSampling
14.7Computer-IntensiveStatistics
14.7.1Cross-validation
14.7.2JackknifeandBootstrapResampling
14.8PickingtheBestModel
Chapter15.Statistics
15.1Data,Counting,andProbability
15.1.1Histograms
15.1.2TypesofCategoricalPredictors
15.1.3Probability
15.1.4Bayes'Theorem
15.1.5Independence
15.1.6CausalityandCollinearity
15.1.7SimplifyingthePredictors
15.2HypothesisTesting
15.2.1HypothesisTestingonaReal-WorldProblem
15.2.2HypothesisTesting,PValues,andAlpha
15.2.3MakingMistakesinRejectingtheNullHypothesis
15.2.4DegreesofFreedom
15.3ContingencyTables,theChiSquareTest,andNoncausalRelationships
15.3.1ContingencyTables
15.3.2TheChiSquareTest
15.3.3SometimesStrongRelationshipsAreNotCausal
15.4Prediction
15.4.1LinearRegression
15.4.2OtherFormsofRegression
15.5SomeCurrentOfferingsofStatisticsTools
15.5.1SASInstitute
15.5.2SPSS
15.5.3MathSoft
Chapter16.ArtificialIntelligence
16.1DefiningArtificialIntelligence
16.2ExpertSystems
16.3FuzzyLogic
16.4TheRiseandFallofAl
Part4.DataMining
Chapter17.IntroductiontoDataMining
17.1DataMiningHasComeofAge
17:2TheMotivationforDataMiningIsTremendous
17.3LearningfromYourPastMistakes
17.4DataMining?Don'tNeedIt--I'veGotStatistics
17.5MeasuringDataMiningEffectiveness:Accuracy,Speed,andCost
17.6EmbeddingDataMiningintoYourBusinessProcess
17.7TheMoreThingsChange,theMoreTheyRemaintheSame
17.8DiscoveryversusPrediction
17.8.1GoldinThemTharHills
17.8.2Discovery--FindingSomethingYouWeren'tLookingFor
17.8.3Prediction
17.9Overfitting
17.10StateoftheIndustry
17.10.1TargetedSolutions
17.10.2BusinessTools
17.10.3BusinessAnalystTools
17.10.4ResearchAnalystTools
17.11ComparingtheTechnologies
17.11.1BusinessScoreCard
17.11.2ApplicationsScoreCard
17.11.3AlgorithmicScoreCard
Chapter18.DecisionTrees
18.1WhatIsaDecisionTree?
18.2BusinessscoreCard
18.3WheretoUseDecisionTrees
18.3.1Exploration
18.3.2DataPreprocessing
18.3.3Prediction
18.3.4ApplicationsScoreCard
18.4TheGeneralIdea
18.4.1GrowingtheTree
18.4.2WhenDoestheTreeStopGrowing?
18.4.3WhyWouldaDecisionTreeAlgorithmPreventtheTreeFromGrowingIfThereWeren'tEnoughData?
18.4.4DecisionTreesAren'tNecessarilyFinishedafterTheyAreFullyGrown
18.4.5AretheSplitsatEachLeveloftheTreeAlwaysBinaryYes/NoSplits?
18.4.6PickingtheBestPredictors
18.4.7PickingtheRightPredictorValuefortheSplit
18.5HowtheDecisionTreeWorks
18.5.1HandlingHigh-CardinalityPredictorsinID3
18.5.2C4.5EnhancesID3
18.5.3CARTDefinition
18.5.4PredictorsArePickedasTheyDecreasetheDisorderoftheData
18.5.5CARTSplitsUnorderedPredictorsbyImposingOrderonThem
18.5.6CARTAutomaticallyValidatestheTree
18.5.7CARTSurrogatesHandleMissingData
18.5.8CHAID
18.6CaseStudy:PredictingWirelessTelecommunicationsChumwithCART
18.7StrengthsandWeaknesses
18.7.1AlgorithmScoreCard
18.7.2StateoftheIndustry
Chapter19.NeuralNetworks
19.1WhatIsaNeural'Network?
19.1.1Don'tNeuralNetworksLearntoMakeBetterPredictions?
19.1.2AreNeuralNetworksEasytoUse?
19.1.3BusinessscoreCard
19.2WheretoUseNeuralNetworks
19.2.1NeuralNetworksforClustering
19.2.2NeuralNetworksforFeatureExtraction
19.2.3ApplicationsScoreCard
19.3TheGeneralIdea
19.3.1WhatDoesaNeuralNetworkLookLike?
19.3.2HowDoesaNeuralNetworkMakeaPrediction?
19.3.3HowIstheNeuralNetworkModelCreated?
19.3.4HowComplexCantheNeuralNetworkModelBecome?
19.3.5HiddenNodesAreLikeTrustedAdvisorstotheOutputNodes
19.3.6DesignDecisionsinArchitectingaNeuralNetwork
19.3.7DifferentTypesofNeuralNetworks
19.3.8KohonenFeatureMaps
19.3.9HowDoestheNeuralNetworkResembletheHumanBrain?
19.3.10ANeuralNetworkLearnstoSpeak
19.3.11ANeuralNetworkLearnstoDrive
19.3.12TheHumanBrainIsStillMuchMorePowerful
19.4HowtheNeuralNetworkWorks
19.4.1HowPredictionsAreMade
19.4.2HowBackpropagafionLearningWorks
19.4.3DataPreparation
19.4.4CombattingOverfitting
19.4.5ApplyingandTrainingtheNeuralNetwork
19.4.6ExplainingtheNetwork
19.5CaseStudy:PredictingCurrencyExchangeRates
19.5.1TheProblem
19.5.2Implementation
19.5.3Theresults
19.6StrengthsandWeaknessess
19.6.1AlgorithmScoreCard
19.6.2SomeCurrentMarketOfferings
19.6.3Radial-Basis-FunctionNetworks
19.6.4GeneticAlgorithmsandNeuralNetworks
19.6.5SimulatedAnnealingandNeuralNetworks
Chapter20.NearestNeighborandClustering
20.1BusinessScoreCard
20.2WheretoUseClusteringandNearest-NeighborPrediction
20.2.1ClusteringforClarity
20.2.2ClusteringforOutlierAnalysis
20.2.3NearestNeighborforPrediction
20.2.4ApplicationsScoreCard
20.3TheGeneralIdea
20.3.1ThereIsNoBestWaytoCluster
20.3.2HowAreTradeoffsMadeWhenDeterminingWhichRecordsFallintoWhichCluster
20.3.3ClusteringIstheHappyMediumbetweenHomogeneousClustersandtheLowestNumberofClusters
20.3.4WhatIstheDifferencebetweenClusteringandNearest-NeighborPrediction?
20.3.5WhatIsann-DimensionalSpace?
20.3.6HowIstheSpaceforClusteringandNearestNeighborDefined?
20.4HowClusteringandNearest-NeighborPredictionWork
20.4.1Lookingatann-DimensionalSpace
20.4.2HowIs"Nearness'Defined?
20.4.3WeightingtheDimensions:DistancewithaPurpose
20.4.4CalculatingDimensionWeights
20.4.5HierarchicalandNonhierarchicalClustering
20.4.6Nearest-NeighborPrediction
20.4.7KNearestNeighbors--VotingIsBetter
20.4.8GeneralizingtheSolution:PrototypesandSentries
20.5CaseStudy:ImageRecognitionforHumanHandwriting
20.5.1TheProblem
20.5.2SolutionUsingNearest-NeighborTechniques
20.6StrengthsandWeaknessess
20.6.1AlgorithmScoreCard
20.6.2PredictingFutureTrends
Chapter21.GeneticAlgorithms
21.1WhatAreGeneticAlgorithms?
21.1.1HowDoTheyRelatetoEvolution?
21.1.2GeneticAlgorithms,ArtificialLife,andSimulatedEvolution
21.1.3HowCanTheyBeUsedinBusiness?
21.1.4BusinessScoreCard
21.2WheretoUseGeneticAlgorithms
21.2.1GeneticAlgorithmsforOptimization
21.2.2GeneticAlgorithmsforDataMining
21.2.3ApplicationsScoreCard
21.3TheGeneralIdea
21.3.1DoGeneticAlgorithmsGuesstheRightAnswer?
21.3.2AreGeneticAlgorithmsFullyAutomated?
21.3.3CostMinimization:TravelingSalesperson
21.3.4CooperationStrategies:Prisoner'sDilemma
21.4HowtheGeneticAlgorithmWorks
21.4.1TheOverallProcess
21.4.2SurvivaloftheFittest
21.4.3Mutation
21.4.4SexualReproductionandCrossover
21.4.5ExplorationversusExploitation
21.4.6TheSchemaTheorem
21.4.7Epistasis
21.4.8ClassifierSystems
21.4.9RemainingChallenges
21.4.10Sharing:ASolutiontoPrematureConvergence
21.4.11MetalevelEvolution:TheAutomationofParameterChoice
21.4.12ParallelImplementation
21.5CaseStudy:OptimizingPredictiveCustomerSegments
21.6StrengthsandWeaknessess
21.6.1AlgorithmScoreCard
21.6.2StateoftheMarketplace
21.6.3PredictingFutureTrends
Chapter22.RuleInduction
22.1BusinessScoreCard
22.2WheretoUseRuleInduction
22.2.1WhatIsaRule?
22.2.2WhattoDowithaRule
22.2.3Caveat:RulesDoNotImplyCausality
22.2.4TypesofDatabasesUsedforRuleInduction
22.2.5Discovery
22.2.6Prediction
22.2.7ApplicationsScoreCard
22.3TheGeneralIdea
22.3.1HowtoEvaluatetheRule
22.3.2ConjunctionsandDisjunctions
22.3.3Defining"Interestingness"
22.3.4OtherMeasuresofUsefulness
22.3.5RulesversusDecisionTrees
22.4HowRuleInductionWorks
22.4.1ConstructingRules
22.4.2ABrute-ForceAlgorithmforGeneratingRules
22.4.3CombiningEvidence
22.5CaseStudy:ClassifyingU.S.CensusReturns
22.6StrengthsandWeaknesses
22.7CurrentOfferingsandFutureImprovements
Chapter23.SelectingandUsingtheRightTechnique
23.1UsingtheRightTechnique
23.1.1TheDataMiningProcess
23.1.2WhatAlltheDataMiningTechniquesHaveinCommon
23.1.3CasesinWhichDecisionTreesAreLikeNearestNeighbors
23.1.4RuleInductionIsLikeDecisionTrees
23.1.5CouldYouDoLinkAnalysiswithaNeuralNetwork?
23.2DataMiningintheBusinessProcess
23.2.1AvoidingSomeBigMistakesinDataMining
23.2.2UnderstandingtheData
23.3TheCaseforEmbeddedDataMining
23.3.1TheCostofaDistributedBusinessProcess
23.3.2TheBestWaytoMeasureaDataMiningTool
23.3.3TheCaseforEmbeddedDataMining
23.4HowtoMeasureAccuracy,Explanation,andIntegration
23.4.1MeasuringAccuracy
23.4.2MeasuringExplanation
23.4.3MeasuringIntegration
23.5WhattheFutureHoldsforEmbeddedDataMining
Part5.DataVisualizationandOverallPerspective
Chapter24.DataVisualization
24.1DataVisualizationPrinciples
24.2ParallelCoordinates
24.3VisualizingNeuralNetworks
24.4VisualizationofTrees
24.5StateoftheIndustry
24.5.1AdvancedVisualSystems
24.5.2AltaAnalytics
24.5.3BusinessObjects
24.5.4IBM
24.5.5PilotSoftware
24.5.6SiliconGraphics
Chapter25.PuttingItAllTogether
25.1DesignforScalability
25.2DataQuality
25.3ImplementationNotes
25.3.1OperationalDataStores
25.3.2DataMarts
25.3.3StarSchema
25.4MakingtheMostofYourWarehouse
25.5TheDataWarehousingMarket
25.6CostsandBenefits
25.6.1BigData--BiggerReturns
25.6.2LawofDiminishingReturns
25.7AUnifyingViewofBusinessInformation
25.8What'sNext
25.8.1DistributedWarehouseEnvironments
25.8.2UsingtheInternetorIntranetforInformationDelivery
25.8.3Object-RelationalDatabases
25.8.4VeryLargeDatabases(VLDBs)
25.9Conclusion
AppendixA.Glossary
AppendixB.BigData--BetterReturns:LeveragingYourHiddenDataAssetstoImproveROI
AppendixC.Dr.E.F.Codd's12GuidelinesforOLAP
AppendixD.10MistakesforDataWarehousingManagerstoAvoidBibliography605
Index609