联合数字图书馆信息服务的关键技术研究

《联合数字图书馆信息服务的关键技术研究》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、中国科学院计算技术研究所博士学位论文联合数字图书馆信息服务的关键技术研究姓名：欧洁申请学位级别：博士专业：计算机应用技术指导教师：林守勋2002.6.1壁鱼墼兰鬯!!堕笪星堡鱼堕茎丝彗』塑墅羔——型婴翌兰—————————————一ResearchonKeyTechniquesofInformationSerVicesinFederatedDigitalLibrarvOuJie(AppliedcomputerTechnology)DirectedByLinShouxunTheinfomationsen，icesinfedera

2、tedd培itallibmryad出essedinmispaperareassociatedwi也thestorage，retrieval，discovery，andusageofenormousdistributcdd电italobjectsontheIntemetreasonablya11de仃ectiVely．Inthisthesis，meinfhsnllctureofinfo舯ationservicesinfederateddigitallibmryispresen协d．Itiscomposedofmesetofsen

3、，icesthatarenecessarytopmvidebasicdigitaIlibrarvfunctiOnali饥whicharecalledcoreservices．Asanin疗．astlllctureofinfonIlationservicesinfederateddigi协llibrary，也edigitalobjectarchitectureisdesc曲ed，whichproVidesmemechallismforpermanentstomge，secureaccess，mallagementandindex

4、todigitalobjects，Thereforethedis仃ibutedenormousdigitalresollrcesinfederateddigitallibrarycanbetrallsfo珊edimoknowledgeforusers．T11einfbnnationre仃ievaljndj垂talobjec￡architectIlrcjsachallen昏ngtopic，whosen．endissem枷cindextofcderatedaIIddis硒butedresourcesbasedonAItechnol

5、o科．SomekeyteclHliquesforthe仃endarcprcsentedinthisp印er，includingsearchengineselectioninfederatedinfo彻ationretrieVal，inteIligentinfonnationretrieVaIbasedonBayesiannetworkmodel，andperSonalizedinfomationservices．Themajncontrjbutionsaredesc“bedasf．ollows：(1)Tbesemallticm

6、ea力jngs砌0ngthete舢saremjnedby瑚eansofassocjatjonru】esdiscoVeryandfoureffectiVemethodsofcomputingtheconditionalprobabilitiesamongte咖sarcpmposed．ThesemanticmeaningsarestoredintheBayesiannetworkmodel，so协einfonnationretrievalsvstembasedonBavesiannetworkmodelcanaccuratelvs

7、earchinformationaccordingtothesemaIlticmeaningofuser’squery．(2)AstatisticalmethodbasedonrelevancetemlstodescriberesourceisDutfonvard，whichimproVesthemethodbasedonthehypothesisofindependenceamongtems．Therefore，theaccuracyofsearchengineselectioninfederatedinfomationre

8、trievalisimprovedandmecon叩lexi哆ofcomputationtoselectthesearchenginesisdecrcased．Themethodisimplementedasf01lows：aRertherelevancete咖saremin