基于數(shù)據(jù)庫的數(shù)據(jù)挖掘技術(shù)的研究及其在工業(yè)中的應(yīng)用.pdf

1、東北大學(xué)碩士學(xué)位論文基于數(shù)據(jù)庫的數(shù)據(jù)挖掘技術(shù)的研究及其在工業(yè)中的應(yīng)用姓名：齊洪鋼申請學(xué)位級別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用指導(dǎo)教師：王大玲2002.2.1盔a盤堂塑堂焦逢塞一—笪四衛(wèi)l叢強(qiáng)StudyandApplicationinIndustryofDataMiningonDatabaseAbstractRecentlythedataminingondatawarehouseanddatabaseasacomputerapplicationdi

2、rectionisdeveloping，F(xiàn)orsolvingthecontradictionbetweenexcessivedataandpoorknowledgenowadays，immensedataaregotfromsubjectdatastoredinwarehouseordatastoredindatabase，andrefined，potential，unknownandusefulknowledgeCanbeextrac

3、tedfromthedatathroughcombiningmathematics，computeranddomainknowledgetogetherThispaperpicksoutproperapproachtoapplydataminingondatabasetoproductionpracticeinindustryenterprisesdependentonthestudyindataminingondatawarehous

4、eanddatabase，andbuildsonlinedatacollectingsystemandoffiinedataanalysissystemWiththeoffiinesystem，knowledgeisgainedthroughanalyzingandminingthedatacollectedfromonlinedatacollectingsystem，andisappliedtoproductionpracticeRe

5、gressionincludingevidencefactoranalysisandprincipalcomponentanalysisisusedindataminingTheevidencefactorswhichinfluencetheproductqualityaregainedwimtheevidencefactormothodTheprincipIalcomponentsofevidencefactorswhicharcor

6、thogonalandlinearindependenceeachotheralegainedwithprincilclalcomponentmethodThenumberofprincipalcomponentsisfewerthanthatofevidencesDecisiontreeanalysisclassifiestheanalyzeddata，findsoutclassifiedfeaturesbetweendifferen

7、tclassesevaluatestheelassifiedfeaturesgenerallythengainsthedecisioninformationtoimprovetheproductqmlity111eprincipalcomponentanalysisgainstheprincipalcomponentstakingplaceoftheevidencefactorsastheartificialneuralnetwork’