基于貝葉斯k近鄰和主成分分析的教務(wù)數(shù)據(jù)挖掘研究.pdf

1、Y12兒^3S分類號(hào)工學(xué)碩士學(xué)位論文(高校教師)密紐編號(hào)基于貝葉斯k近鄰和主成分分析的教務(wù)數(shù)據(jù)挖掘研究碩士研究生指導(dǎo)教師學(xué)科、專業(yè)學(xué)位論文主審人陳秀玲李智慧副教授計(jì)算機(jī)應(yīng)用技術(shù)吳良杰教授哈爾濱X程大學(xué)2007年6月哈爾濱工程大學(xué)碩士學(xué)位論文AbstraotUniversityeducationaladministrationmanagementisanextremelyimportantlinkinhi【ghereducationIti

2、sthecoreandfoundationofentireuniversitymanagementAgeneralviewofformerteachingmanagementsystemsshowsthatmostofthemareOnLineAnalyticalProcessing(OLAP)systemwhichlacktheabilitiesofcomprehensiveanalysisandassistantdecisionma

3、kingFurthermoretheyfailedtousetheimplicitknowledgeintheinformationwithovergreatcapacityaccumulatedduringthehistoryHeretheeducationaladministrationdataminingisimplementedbymakinguseofthedataminingtheoryInthisthesistheelem

4、entarytheoryofdataminingtechnologyandthemainalgorithmsarediscussedfirstlyResearchingstatusofdataminingisreviewedandsummarizedTheapplicationsofprincipalcomponentanalysisandBayesianknearestneighboralgorithmindataminingares

5、tudiedInthethesisprinciplecomponentanalysisisusedinsyntheticalevaluationofthegraduatingstudents’achievementByremovingthecorrelationbetweenthefactorstheanalysistargetsarereducedwhiletheamountofinformationiskeptSOthattheco

6、mputationisdecreasedThethesisadoptsprincipalcomponentanalysisandBayesianknearestneighhonralgorithminforecastofstudents’futureoccupationsnestudents’achievementsareusedasfeaturedataandprincipalcomponentanalysisisusedteduce

7、thedimensionsoffeaturedataBayesianknearestneighbouralgorithmisappliedtoimplementclassification，namelypredictionoffutureoccupationsComparedwithtraditionalknearestneighbouralgorithmBayesianknearestneighbouralgorithmisablet

8、odeterminetheparametersofalgorithmbyBayesianlearningandMarkovChainMonteCarloalgorithm(MCMC)SothoseparametersneednottobeassumedasbeforeandtheresultsofclassificationaresteadierthanbeforeThesimulatedexperimentresultsproveth