05CubeTech.pptVIP免费

1DataMining:ConceptsandTechniques(3rded.)—Chapter5—JiaweiHan,MichelineKamber,andJianPeiUniversityofIllinoisatUrbana-Champaign&SimonFraserUniversity©2011Han,Kamber&Pei.Allrightsreserved.201/26/25DataMining:ConceptsandTechniques23Chapter5:DataCubeTechnologyDataCubeComputation:PreliminaryConceptsDataCubeComputationMethodsProcessingAdvancedQueriesbyExploringDataCubeTechnologyMultidimensionalDataAnalysisinCubeSpaceSummary4DataCube:ALatticeofCuboidstime,itemtime,item,locationtime,item,location,suppliercalltimeitemlocationsuppliertime,locationtime,supplieritem,locationitem,supplierlocation,suppliertime,item,suppliertime,location,supplieritem,location,supplier0-D(apex)cuboid1-Dcuboids2-Dcuboids3-Dcuboids4-D(base)cuboid5DataCube:ALatticeofCuboidsBasevs.aggregatecells;ancestorvs.descendantcells;parentvs.childcells1.(9/15,milk,Urbana,Dairy_land)2.(9/15,milk,Urbana,*)3.(*,milk,Urbana,*)4.(*,milk,Urbana,*)5.(*,milk,Chicago,*)6.(*,milk,*,*)alltime,itemtime,item,locationtime,item,location,suppliertimeitemlocationsuppliertime,locationtime,supplieritem,locationitem,supplierlocation,suppliertime,item,suppliertime,location,supplieritem,location,supplier0-D(apex)cuboid1-Dcuboids2-Dcuboids3-Dcuboids4-D(base)cuboid66CubeMaterialization:FullCubevs.IcebergCubeFullcubevs.icebergcubecomputecubesalesicebergasselectmonth,city,customergroup,count(*)fromsalesInfocubebymonth,city,customergrouphavingcount(*)>=minsupportComputingonlythecuboidcellswhosemeasuresatisfiestheicebergconditionOnlyasmallportionofcellsmaybe“abovethewater’’inasparsecubeAvoidexplosivegrowth:Acubewith100dimensions2basecells:(a1,a2,….,a100),(b1,b2,…,b100)Howmanyaggregatecellsif“havingcount>=1”?Whatabout“havingcount>=2”?icebergcondition7IcebergCube,ClosedCube&CubeShellIsicebergcubegoodenough?2basecells:{(a1,a2,a3...,a100):10,(a1,a2,b3,...,b100):10}Howmanycellswilltheicebergcubehaveifhavingcount(*)>=10?Hint:Ahugebuttrickynumber!Closecube:Closedcellc:ifthereexistsnocelld,s.t.disad...