基于Web挖掘的决策支持系统模型研究

《基于Web挖掘的决策支持系统模型研究》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、一lnDⅧ061⑧天弗大薯十目*代第一"★{博士学位论文___■疆●■_嵋■跚j：■■■■●¨一级学科：筐垄哇主与工捏学科专业：筐理型生与王拦作者姓名：衄厦指导教师：鞋室垂越撞天津大学研究生院2005年I2月ABSTRACTAccompanyingthedevelopmentofinformationtechnologyandE-commerce，themodelofknowledgeanddecisionsupportunderthenetworkhasbeenbecominganewhotspottopicinthedomainsofE-com

2、merce，managementscienceandartificialintelligenceetc．Intheinformationsocietywiththenetworkeconomydevelopingrapidly，itisveryimportantbothinscienceandinapplicationtomakeathorou【ghstudyonthemodelsofknowledgediscoveryanddecisionsupportbasedonthewebmining．Themaincontentsdiscussedint

3、hisdissertationareasfollows：Byreferencingtoexistingliteratureinrelatedfields，theprocess，methods，functions，themissionsandtargetsineachphaseoftheprocessmodelofdataminingaresummarizedandreviewed，theconcept，categoryandmaintasksofwebminingareexpounded．Thewebdocumentsclassificationm

4、odelbasedonconceptfeaturevectorsarebuilt．Inordertomanageandclassifythesemi—structurewebdocuments，itadvancestheconceptofconceptfeaturevectors，describesthecollectionofconceptfeaturewordsandthecalculationmethodofWeights，constructstheNBandAssociationRulesclassifiersbasedonconceptf

5、eaturevectors，andverifiesthevalidityandcorrectnessofthewebdocumentsclassificationmodelbasedonconceptfeaturevectorsbyexperiments,Themodelsofknowledgediscoverybasedonthemanagementofwebusagedataaresetup．ItexpoundstheURL-UserlDassociatedmatrixexpressmethodforbrowsingbehaviorsofweb

6、users，establishesthemodelsofwebpagesclusteringandwebconsumersclusteringbasedonURL-UserlDassociatedmatrix，constructsthemulti—Markovchainspredictionmodelofwebuserbrowsingbehaviorsbasedonclusteredwebusers，andthemethodoftheminingmodelofmaximalfrequentsequencesbasedonclusteredwebus

7、ersandSUffixtrees．Themodelsofknowledgediscoveryundertheenvironmentofsemanticwebarediscussed．Itintroducesbackground，thesystemstructureofthesemanticwebandwebontology，elaboratestheexpressionmodelofdomainontology,discussestheclassificationprocessofwebpagesunderthesemanticweb，build

8、stheclassificationmodelofwebpagesbasedonthedomainontologyandt