案例推理在决策支持系统中的应用研究

《案例推理在决策支持系统中的应用研究》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案例推理在决策支持系统中的应用研究姓名：王津津申请学位级别：硕士专业：计算机系统结构指导教师：田卫东20100401ResearchonCase—-BasedReasoninginDecisionSupportSystemAbstractCase-BasedReasoning(CBR)isanimportantmethodinsolvingproblemintheareaofartificialintelligence，itshowsgreatvitalitybecauseofi

2、t’SspecialreasoningstyleandSuccessapplication，moreandmorepeopleshowwiderangeofinteresttoCBR．CBRcanbeunderstoodasamethodofusingthepastcasesandexperiencetosolvenewproblems，italsocanbeconsideredasananalogyinferencefromonecasetoanotherone．Withthedevelopmento

3、fcomputerscience，usingthecomputerfordecisionsupportisbecomingatrend．IntheviewofdesigningthemoduleofCBRindecisionsupportsystem(DSS)，thisdissertationresearchedthekeytechnologyofCBRsuchascaserepresentation，casebase，caseretrieval．Themaincontentsareinbriefasf

4、ollowing：(1)Onthebasisoftheintroductiontotheorigins，developmentandresearchstatusofCBR，thebasicmethod，systemframeworkandthekeytechnologyofCBRarediscussedindetailinthisdissertation．(2)Accordingtotheproblemofdesigningthecomplicatedtemporalcasesandeasebasein

5、somespecificareas．Basedonusingframeworkasthebasicmethodforrepresentingcase，wedescribedthecasewiththeattributeoftime；Itcanreducethecasebase’Ssizeintheconditionofmaintainthecrucialinformationofcasesbyusingcorrectionanalysistoselectfeatureattributeandprinci

6、palcomponentsanalysistorebuildfeatureattribute．(3)Forthepurposeofraisingtheefficiencyofcaseretrieve，firstlyinductionindexmethodisusedforreducingthesearchingspace，andthennearestneighbormethodisusedforsearchingthemostsimilarcase．Inordertosolvingtheproblemo

7、fsettingthefeatureweightwhencaseiSretrieved，acombinedmethodforsettingfeatureweightbasedonentropywas．proposedinthisdissertation．Themethod’Svalidityistestifiedbyexperiment．(4)TheimplementationofCBRmoduleindisasterrescuedecisionsupportsystemwasstudiedfinall

8、y．Keywords：Case-BasedReasoning，DSS，easebase，caseretrieval，entropy，combinedweightU插图清单图2．1案例推理系统框架⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯9图3．1框架表示图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯一15图3．2属性相关性⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯17图4。1案例检索策略⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯