[1] Mining Association Rules between Sets of Items in Large Databases.pdf

《[1] Mining Association Rules between Sets of Items in Large Databases.pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、MiningAssociationRulesbetweenSetsofItemsinLargeDatabasesRakeshAgrawalTomaszImielinskiArunSwamiIBMAlmadenResearchCenter650HarryRoad,SanJose,CA95120AbstractSeveralorganizationshavecollectedmassiveamountsofsuchdata.ThesedatasetsareusuallystoredWearegivenalargedatabaseofcustomertransactions.ontert

2、iarystorageandareveryslowlymigratingtoEachtransactionconsistsofitemspurchasedbyacustomerdatabasesystems.Oneofthemainreasonsfortheinavisit.Wepresentanecientalgorithmthatgeneratesalllimitedsuccessofdatabasesystemsinthisareaissignicantassociationrulesbetweenitemsinthedatabase.thatcurrentdatabase

3、systemsdonotprovidenecessaryThealgorithmincorporatesbuermanagementandnovelestimationandpruningtechniques.Wealsopresentresultsfunctionalityforauserinterestedintakingadvantageofapplyingthisalgorithmtosalesdataobtainedfromaofthisinformation.largeretailingcompany,whichshowstheeectivenessoftheThis

4、paperintroducestheproblemofmining"alargealgorithm.collectionofbasketdatatypetransactionsforassocia-tionrulesbetweensetsofitemswithsomeminimumspeciedcondence,andpresentsanecientalgorithm1Introductionforthispurpose.AnexampleofsuchanassociationruleConsiderasupermarketwithalargecollectionofitem

5、s.isthestatementthat90%oftransactionsthatpurchaseTypicalbusinessdecisionsthatthemanagementofthebreadandbutteralsopurchasemilk.Theantecedentsupermarkethastomakeincludewhattoputonsale,ofthisruleconsistsofbreadandbutterandthecon-howtodesigncoupons,howtoplacemerchandiseonsequentconsistsofmilkalone.

6、Thenumber90%istheshelvesinordertomaximizetheprot,etc.Analysiscondencefactoroftherule.ofpasttransactiondataisacommonlyusedapproachTheworkreportedinthispapercouldbeviewedasainordertoimprovethequalityofsuchdecisions.steptowardsenhancingdatabaseswithfunctionalitiesUntilrecently,however,onlyglobal

7、dataaboutthetoprocessqueriessuchas(wehaveomittedthecumulativesalesduringsometimeperiod(aday,aweek,condencefactorspecication):amonth,etc.)wasavailableonthecomputer.Progressinbar-codetechnologyhasmadeitpossibletostoretheFindallru