基于归纳学习和范例推理的智能决策支持系统的研究

《基于归纳学习和范例推理的智能决策支持系统的研究》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、硕士学位论文学校代码：10357学号：论文题目(中文)基于归纳学习和范例推理的智能决策支持系统的研究TheResearchoflDSSBasedonlnductiveLearningandCase—BasedReasonIng论文题目(英文)姓名学科专业研究方向指导教师完成时间李学俊计算机应用机器学习与智能软件倪志伟2005年5月慕于归纳学习和范例推理的智能决策支持系统的研究AbsIracITHERESEARCH0FIDSSBASEDONINDUCTIVELEARNINGANDCASE—BASEDREASONINGABSTRACTDecisionSupportSystemisi

2、ntegraledwithAftificialInlelligencetOconstiIuteIntelligenceDecisionsupporcSyslem(IDss)．IDsscane虢ctiVelysolVemeprobIemofhowqualitativeanalysisiscombinedwithquantitatiVeanalysisandthosestructuredandnon—structuredproblems，extenditsscopewhichitcandealwith，andimproVethedecisionability．IDSShasbeen

3、widelyappliedtosomefields，suchaspreventingnood，supportinge—business．InductiVeL_eamingandCase—basedReasoning(cBR)arebothkeytechniquesofmachinelearningwhichhasfullybeendeveloped，andIheyhavebeenwidelyresearchedandapplied．HoweVerwjlh【heappJicationofIDSS，theproblemof‘‘theboItlencckofacqui“ngknowl

4、edge”has争aduallycomeoutandhinderediIsdevelopmentinsomeexlent．Fortunatelythetwokeytechniquescanapp】iedtoIDSStoe凰ctiVelysolVetheproblemmentioned¨boVe，soastoeffccfivelyande所cientlysupportthedecision．Thisthesisfirstlydiscussesmeprinciple，architectureconstruction，function，reseafchsiIuationandexis

5、tingprob】emsofIDSS，The8pplicaIionofMachineLearningisresearched【osolveiheproblemof“thebottlcncckofacqujrjngknowledge’’IhatexistsinthestudyofIDSS．AnditinlfoducessomeaIgo“lhmsofdecisiontreeleamingsuchasID3，c4．5andfeaIuresubsetseJectionofInducliVeL￡aming．Thenitalsointroducessomebasictechnjquesof

6、CBRsuchascaserepresentation，caseretricval，casereVisal，casebasemaintenance．Acco『dingtothesetheoriesmentionedabove，thethesisraisessomenewapplicationofInductiveLearninginCBR．SomealgorithmsofDecisiontreelearning，especiallyC4．5，areappliedtocaseretrieval，casereVisalandcasebasemainIenance．Ⅱ幕于归纳学习和范

7、例推理的智能决策支持系统的研究AndaIsoseVefalaJgo“

8、hmsoffcaluresubsetseleclion，especial】yfhealgorilhmwhichisbasedontheinformationentropyandLaplaceerrorevaluationfuncIionsandabletodealwithnoisydata，areappliedtothereductionofattributesofcaseinthecasebase．FinaJly，com