欢迎来到天天文库
浏览记录
ID:37352464
大小:2.84 MB
页数:78页
时间:2019-05-22
《基于数据仓库的知识发现模式的研究》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、合肥工业大学硕士学位论文基于数据仓库的知识发现模式的研究姓名:马玉宝申请学位级别:硕士专业:计算机软件与理论指导教师:胡学钢20030401合肥T业人学硕i:iq文摘要AbstractItiSahotspotthatthedataminingoftimeserialmodel.classifyrule.associationruleinthedataminingstudycurrently.Inthisthesis,thethoroughstudyoftimeserialmodel.classificationnileandass
2、ociationruleiSmade.ItiSthoughttllatthedataminingisthemultistageprocessofriser’scenterinthisthesis.ItiSimportantthatthedataminingofmultistageprocessshouldbestudyintheeachphase.Inthisthought.themainworkinthisthesisasfoliowing:ThewayofthedataminingiSdeeplystudiedandamode
3、lofdataminingbasedondatawarehouseiSintroduced.ThemodeliScombinedwiththestrate州ofdatawarehouseandOLAPtechnology,madeuseofconcepthierarchyasbackgroundknowledge.andextendeddatainthedatabaseasuser’Sinterestingmetadataoftheconcepthierarchy,aggregatton.Besides.relativityana
4、lysesiSintroducedintheprocessofdatapretreatmentinthisthesis.therebycanceledthedisrelatedattributeofdataminingassignment,reducedIotsofdatasetsandimprovestheaccuracyandefficiencyofrulesmined.Tbbedealedagaistextendeddata。thisthesishasimprovedonandcometruearithmeticoftime
5、sequencemodel,andamendedconventionaldecisiontreearithmetic,introducedthedecisiontreearithmeticforextendeddata,namelythresholdvalueCOntrolapproach.AccordingtOthresholdvalueandconcepthierarchy,thresholdvaluecontrolapproachcansetuptheconciseandstatisticclassificationtree
6、.Atthesametime,basedonthetheoryoftheconceptlattice,thisthesisinttoducesthearithmeticofminingassociationrulesbasedonquantifiedconceptlaaicereducedbyuncertaintycoemcient.Thisthesisdesignsandaccidencilycarriesouttheprototypesystemofdatamining.T11iSt001iScenteredontheuser
7、,undertheuser’Scontr01.andtobecapabletoeriectivelyminetheruleoftimesequencemodelandtheclassificationroleandtheassociationruleinthedatabaseordatawarehouse.Keywords:DataMining,DataWarehouse,OLAP,TimeSequenceModel,ClassRule,ConceptLattice,AssociationRule11独创性声明本人声明所呈交的学位
8、论文是本人在导师指导下进行的研究工作及取得的研究成果。据我所知,除了文中特别加以标注和致谢的地方外,论文中不包含其他人已经发表或撰写过的研究成果,也不包含为获得盒蟹王些太堂或其他教育机构的学位或证书而使用过的材料。与我一同工作的同志对本研究所做的任
此文档下载收益归作者所有