XML模糊結(jié)構(gòu)查詢中檢索結(jié)果的自動(dòng)選取.pdf

1、江西財(cái)經(jīng)大學(xué)碩士學(xué)位論文XML模糊結(jié)構(gòu)查詢中檢索結(jié)果的自動(dòng)選取姓名：蘭月華申請學(xué)位級別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：萬常選20091201AbstractAsaselfdescribingandextensiblelanguage，XMLhasbecomeformatsandstandardsforinformationrepresentationanddataexchange，SOthatithasbeenusedwidelya

2、ppliedtovariousfieldsTherefore，therehavegreatprospectsforstudyingtheaccurateretrievalofXMLdocumentsThegranularityoftheresultsiSthebiggestdifferencebetweenXMLDocumentandtraditionaltextretrievalInthe仃aditionaltextretrieval

3、，thegranularityistheentiredocumentthenusersshouldneedtosearchthewholedocumentagaintogettheinformationtheyreallyneedThereforetheretrievalprecisionisalwayslowHoweverinXMLdocumentretrieval，thegranularityoftheresultsisXMLdoc

4、umentfragment，insteadoftheentiredocumentInthiswayitcarlimprovetheretrievalprecisionRetrievalresultsinXMLqueriesareasetofsubtreescontainingkeywordsandthekeyistochoosetherootnodeoftheresultsub—treewhichiStheanswernodeBecau

5、setheXMLdocumenthasseveralelementsofdifferentgranularitiesSOweshouldtochoosewhichnodesasanswernodesItbecomesoneofthekeyissuestohowtochoosereasonableanswernodesUnliketraditionaltextdocument，XMLdocumentiSSemi—stanacturedda

6、taandspecifiesthecontentsaswellasthestructureinformationItCanberetrievedbymeansofnotonlycontentonlyqueriesbutalsocontentandstructurequeriesContentonlyqueriescontainonlyseveralsimplekeywordsSOtheydonotindicatetheanswernod

7、es；IntheCASquerytheanswernodesmaybenotspecified，orifspecified，theyareoftenvague，inaccurateThence，weneedtoinfertheanswernodesfromtheuserqueryexpressionsTheelementnodesintheXMLdocumentCanbecategorizedintofourgroups：entityn

8、odes，connectionnodes，leafnodesandvaluenodesTakingthesemanticintegrityoftheanswernodesintoaccount，thecandidateanswernodesonlycontainsentitynodesandconnectionnodesBasedonthenodes’classification，wecalculatetheconfidenceofea

9、chcandidateanswernodetobetheanswernodewithrespecttothegivenkeywordqueryThentheoneswiththetopnconfidencearechosenastheanswernodesbutnodesshouldberemovedwhichiStheancestorofothernodeThedepthofnodesandkeywordsmatchingshould