數(shù)據(jù)挖掘中關(guān)聯(lián)規(guī)則的優(yōu)化.pdf

1、中山大學碩士學位論文數(shù)據(jù)挖掘中關(guān)聯(lián)規(guī)則的優(yōu)化姓名：楊敏申請學位級別：碩士專業(yè)：計算機軟件與理論指導教師：印鑒20040508一莖堡絲塑!莖壁塑型竺垡些Title：AssociationRules’OptimizationinDataMiningMajor：ComputerSoftwareandTheoryName：YangMinSupervisor：Vice—ProfessorYinJianABSTRACTTodayDataMining

2、isoneofthemostpopulartechnologiesofcomputerscienceAsanewtechnologythereexistmanyproblemsaboutdataminingtoberesearched。HowtoUSeitcorrectlyHowtomakeanapplication，andHowtoimproveit，alltheseproblemsattractcomputerscientistIn

3、thispaperItrytoinvestigatetheoptimizationoftheassociationrulesinDataMiningonanewpointofview：StrerIgthenthedefinitionofsupportandconfidenceoftheassociationrules，SOthatwedon’tneedtoaddsomenewthresholdtominingassociationrul

4、eswithoutredundantandfalsehoodAlgorithmsinexistenceaboutassociationrulesdiscoveryadoptfrequentKitemsettocreateK十1candidateitemset。Ifthebusinessdatabaseisverybig，thetimeofcountingfrequentitemsetwillbecomparativelylong，and

5、ifwereducethesupportandconfidence，thenumberofassociationruleswillincreaserapidlywithamassofredundanifiesThiswilltakedowntheefficiencyoftotalalgorithmAtthesametime，becauseoftherestrictionofassociationrulesdefiniton，quitea

6、bitofrulesarefalseandmeaninglessThispaperpresentsanewdefinitionofassociationrulesandgivesanalgorithmofminingassociationrulesThisalgorithmasenegativeitemtostrengthenthedefinitionofsupportandconfidence，anddeleteredundancie

7、sofpositiveandnegativerulesbythedefinitionofsimpleredundantruleandstrictredundantruleItadvancethequantityofassociationrulesinintegerSOthattheyareallsignificativekeywords：dataminil】g，associationrule，negativeassociationrul