面向Semantic Web的自動(dòng)語(yǔ)義標(biāo)注方法研究.pdf

1、中國(guó)科學(xué)技術(shù)大學(xué)碩士學(xué)位論文面向SemanticWeb的自動(dòng)語(yǔ)義標(biāo)注方法研究姓名：林樂(lè)申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用指導(dǎo)教師：陳恩紅20040501中罔科學(xué)技術(shù)人學(xué)壩Jj論義嘶向ScmamicWeb的口曲l矗義h、往山；j，叭咒AbstractWiththedevelopmentofIntemetWoridWideV，ebhasbecomeoneofthemainsourcesfromwhichforpeopletogetinfor

2、rnationButbecauseWebinfo咖ationcannotbe“understood”bymachinetofindtheexactinformationforuserrapidlyiSmoreandmoredifficultToslovethisproblemtechnologyofSemanticWebiSdevelopedItiSanextensionlayertothecurrentwebAndonthislaye

3、rwcbinformationcanbcdescribed：。o／lalsemanticsthroughaformofannotationAllexistingannotationmethodneedhuman’SassistantnowHoweverthereiSSOmuchlrlformationorlthewebtodosemanticannotationonlybyhumaniSobviouslynotagoodwaylnthi

4、sarticlewewillintroduceandanalysetheactualityoftheSemanticWebandannotationtechnologyAndthenmakeemphasesonhowtOtealizeautomaticWebinformationannotataionbydataminingmethodThemajortasksandinnovationsarc：1)Totheproblemofauto

5、maticsemanticannotation，designanautomaticsemanticannotationframeworktowardSemanticW曲whichhasfourmainmodules：ontologyeditingandmaintaining，instanceextraction，semanticannotationandqueryAndthecorepartiSinstanceextractionOnt

6、ologyalsohasimportantpositionthroughoutthewholeprocessofsemanticannotaion21TotheproblemofinstanceextractionpresentanontologybasednaiveBeyesclassificationmethodInthismethodaWjbPagerepresentsaclassinstanceAndwecancreateaⅥ，

7、ebpagesclassifierbyintegratingdomainontologyintotraditionalstatistic，basedBeyesclassificationmethodExperimentsshowthatthiSmethodiSeffective31Totheproblernhowtouseclusteringmethodininstanceextractionwepresentallontologyba

8、sedclusteringinstanceextrationmethod，F(xiàn)irstthedefinitionofweightbetweenontologyandwordsisgiven，Andwithit，anontology—baseddocumentsimilaritycacnlationmethodiSdesignedSecondusingthiscaculationmethodweimprovethetraditionalk—

9、meansclusteringalgorithmExperimentsshowthatimprovedclusteringalgorithmdoesextractclassinstanceseffectively41ToextractWOrdorphrasetypeinstancewepresentalexicon—syntacticpatternlearningmethodLexicon—syntacticpatternisaregu

10、larexpressionwithgooddefinitionWhenapplythisPatteminWebPagesWCcangettheinterestinginstanceinthematchingphraseHeretwokindofleiconsyntaticpatternaregiven：special—wordpatternandiuxtapositionpatternWehavedonesomeexperimentsw

11、ithpartofthesepatterns，andgetasatisfiedresult，5)UsingRDFpresentanembeddedsemanticannotationmethodaccordingtodifferenttypesofinstances——PageTypeInstanceandWbrdTypeInstanceThismethodusesaforillofembeddingRDFstatementsintoH