中文WEB文本傾向性分類研究.pdf

1、北京交通大學碩士學位論文中文WEB文本傾向性分類研究姓名：單大力申請學位級別：碩士專業(yè)：信息網(wǎng)絡與安全指導教師：劉云20071201韭塞窯煎太學亟堂僮監(jiān)塞△曼S!基△￡工ABSTRACTABSTRACT：21thcenturyisthetimeswheninformationincreasesexplosivelyMoreandmoreinformationexistsinthewayofelectro—documentalongwit

2、hthequickdevelopmentofIntemet，andmostofthesedocumentsaleunorderlyDocumentautomaticallyclassifyingtechniquecansolvetheunoderlyinformationproblem，organizeandmanagetheinformationeffectivelyandheiptheusersacquiretheinformati

3、onwhichtheyneedquicklyaccuratelyandcomprehensivelyDocuments’orientationtextclassificationisaresearchhotspotandallimportantaspectofnetworkconsensUSsecuritySomeconventionaldocumentsclassificationsystemsclassifythedocuments

4、bytheircontentsuchas：militaryaffairs，medicines，sportsThesesystemscannotidentifydocuments’orientationInordertomaintainthenetworkssecuritywedeveloptheWebdocuments’orientationclassificationsystemItsmaintaskisanalyzingthecon

5、tentofdocumentstoclassifythembytheirorientationItcallidentifythesupportofsomenetworkconsellSUsandhasallimportanteffectofmaintainingthenetworkconsensussecurity低thesisdiscussesSOmecoretechniquesusedindocumentorientationcla

6、ssificationandanalysestheadvantagesanddisadvantagesofthemItProvideslogicaldocumentsAccordingtotheaccuracyandconvenienceprincipleuseC≠tocompletethetextparticiplemodulebasedonfourwordhashtabledictionaryanditstextparticiple

7、accuracyispast90percentWeusecommendatoryandderogatorycharacterpick—uptechnique，vectorspacemodelstructuretechniqueandSVMtoimplementthesystems’documentsorientationclassificationfunctionanditsclassificationaccuracyispast80p

8、ercentAtlast，Weanalyzetheresultoftextparticipleandtextorientationclassificationand晡ngforwardthedisadvantagesofthesystemandthefutureresearchoftentationKEYWORDS：TextOrientation；TextModel；ParticipleDictionary；ChinesePartici