基于本體的文本分類研究.pdf

1、貴州大學(xué)碩士學(xué)位論文基于本體的文本分類研究姓名：吳曉申請學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：張明義傅家祥20080501貴州大學(xué)碩士論文AbstractTheResearchofTextCIassificationBasedonOntoIogyAbstractWiththerapiddevelopmentofInternetandtheenterpriseintranet，allkindsofelectronicdocument

2、shavebeengrowinggreatlyIthasbecomingdifficultforuserstofmdinformationthatsatisfiestheirneedssoithasbecomeagreatchallengeforinformationscienceandtechnologythathowtoorganizeandprocessthemassivedocumentsquicklyexactlyandful

3、lyThetextclassificationisakeytechnologytoprocessandorganizationmassivetextdata；itcanhelppeopletosolvethedifficul哆ofinformationdisordertoagreatextentandtolocatetherequiredinformationquicklyTherefore，thetextclassificationi

4、sgraduallygettingmoreandmoreimportantInthispaperweintroducetextclassificationanditsrelatedtechnologies，pointouttheflawoftraditionaltextclassificationkeywordsaremutuallyindependent，don’thavethesemanticconnection，whichisop

5、posedwiththefactobviouslyThedocumentsarerelatedtoeachotheronlyiftherealesharedkeywordsinthedocuments，suchassynonym，hypemym，hyponymyeteHoweverthedifficultyliesinthefactthatmostkeywordshavemultiplemeaningontheonehand，andon

6、theotherhand，someconceptscanbedescribedbymorethanonekeywordInordertosolvethisproblemweattempttomatchkeywordswithconceptsofdomainOntologyThusweexpressthedocumentwithConceptVectorSpaceModel(CVSM)TheapproachCallkeepthetexti

7、nformationmostlyandtheprecisionofthetextclassificationiseffectivelyimprovedThemainworksareasfollows1Thekeytechnologiesoftextclassificationareanalysedinthisthesisincludingthedefmitionoftextclassification，featureselection，

8、taxonomicapproach，andperformanceevaluationetcIntheend，wepointouttheexistencequestionsoftraditionaltextclassification2Therearecloserelationsbet、ⅣeentextclassificationandpersonalizedinformationretrievalTheresearchonpersona

9、lizedinformationretrievalisdiscussedandanadjustingalgorithmforUSel“profileispresented3OntologyisintroducedThekeywordsarematchedwiⅡ1conceptsofdomainOntologythusVectorSpaceModel(VSM)ofdocumentistransformedintoConceptVector

10、SpaceModel(CVSM)inordertesolvethesynonymy，hypernym，hyponymyetc4AtextclassificationsystemarchitecturebasedonCVSMispresentedandcarriedontheanalysistoeachmoduleAseriesofsimulationtestsindicatetheprecisionandrecalloftextclas