The CRISP-DM User Guide

1 1 The CRISP-DM User GuideBrussels SIG MeetingPete ChapmanNCR Systems Engineering Copenhagenemail: Objectives and BenefitsnCRISP-DM DeliverablesnCRISP-DM Methodology, Phases and TasksnCRISP-DM User GuidenPossible CRISP-DM Futures3 Objectives and Benefits of CRISP-DMuensure quality of knowledge discovery project resultsureduce skills required for knowledge discoveryureduce costs and timeugeneral purpose ( , stable across varying applications)urobust ( , insensitive to changes in the environment)utool and technique independentutool supportableusupport documentation of projectsucapture experience for reuseusupport knowledge transfer and training4 CRISP-DM DeliverablesuProcess ModeluMethodologyuReference ModeluUser GuideuOutput (Deliverable/Templates)

2 UTool SupportuTool Support DefinitionsuStream LibraryuExperimentationuExperimentation ReportsuCRISP-DM SIG User Feedback5 CRISP-DM MethodologyMappingPhasesGeneric TasksCRISPP rocess ModelSpecialized TasksProcess InstancesCRISPP rocess6 Data Mining ContextsSpecialized TasksGeneric TasksApplication Domains Response Modeling Churn Prediction ..Technical Aspects Missing Values Outliers ..Problem Types Data Description / Summarization Segmentation Concept Description Predictive Modeling Dependency AnalysisTools and Techniques Clementine MineSet Decision Trees ..7 CRISP-DM PhasesDataUnderstandingDataPreparationMo dellingDataDataDataBusinessUnderstanding DeploymentEvaluationDataUnderstandingDat aPreparationModellingDataDataDataBusines sUnderstandingDeploymentEvaluation8 Phases and TasksBusinessUnderstandingDataUnderstand ingEvaluationDataPreparationModelingDete rmine Business ObjectivesBackgroundBusiness ObjectivesBusiness Success CriteriaSituation AssessmentInventory of ResourcesRequirements, Assumptions.

3 And ConstraintsRisks and ContingenciesTerminologyCosts and BenefitsDetermine Data Mining GoalData Mining GoalsData Mining Success CriteriaProduce Project PlanProject PlanInitial Asessment of Tools and TechniquesCollect Initial DataInitial Data Collection ReportDescribe DataData Description ReportExplore DataData Exploration Report Verify Data Quality Data Quality ReportData SetData Set DescriptionSelect Data Rationale for Inclusion / ExclusionClean Data Data Cleaning ReportConstruct DataDerived AttributesGenerated RecordsIntegrate DataMerged DataFormat DataReformatted DataSelect Modeling TechniqueModeling TechniqueModeling AssumptionsGenerate Test DesignTest DesignBuild ModelParameter SettingsModelsModel DescriptionAssess ModelModel AssessmentRevised Parameter SettingsEvaluate ResultsAssessment of Data Mining Results Business Success CriteriaApproved ModelsReview ProcessReview of ProcessDetermine Next StepsList of Possible ActionsDecisionPlan DeploymentDeployment PlanPlan Monitoring and MaintenanceMonitoring and Maintenance PlanProduce Final ReportFinal ReportFinal PresentationReview ProjectExperience DocumentationDeployment9 Introduction to the User GuideGeneric TasksSpecialized TasksContextReference ModelWhat To Do?

4 User GuideHow To Do? check lists questionaires tools sequences of steps decision points pitfalls10 CRISP-DM User Guide11 How to use the User Guide (i)uContents of the User Guide - More detailed description of the various tasks using:uActivities ListuCheck ListsuGood IdeasuWarnings!uWhat is NOT in the User GuideuDeliverables/Document Templates (as yet)uDescription of Techniques and Tools (as yet)uEstimates of engagementsuQuality Indicators12 How to use the User Guide (ii)uBeginning Data MinersuWhat tasks do I need to do?uWhat is the order of the tasks in a Data Mining Engagement?uWhat risks do I run?uAre there any shortcuts in my tasks?uWhat are the format of the deliverables that I need to resent to management?

5 UExperienced Data MinersuHave I missed any activity?uAre there any tasks or activity that I can leave until later?uHow can I make a Project Plan?uHow can I document the project for later re-use?13 Possible Future CRISP-DM Deliverablesu CRISP-DM - The Book , includesuExperiences, feedback from SIG membersuReference Model, User Guide updated with experimentsuFull Deliverables/Document TemplatesuCase StudiesuMapping Advice from Generic to Specific EngagementsuMore explicit advice on Tools & TechniquesuAdvice on documentation of engagements,establishment of Data Mining Library,..14