1、分類號(hào)UDC密級(jí)——編號(hào)魚2【垃Q業(yè)互學(xué)位論文李文，指導(dǎo)教師姓名：王僉延熬援直差主盍堂鹽簋主!墜申請(qǐng)學(xué)位級(jí)別：亟專業(yè)名稱：鹽篡扭鏊佳曼理迨論文提交日期：2Q墮生5旦論文答辯日期：呈堂生魚旦日學(xué)位授予單位和日期：直昌盔堂2壁壁墨生旦目答辯委員會(huì)主席：評(píng)閱人：2005年5月日碩二匕學(xué)位論文：公安執(zhí)法監(jiān)督管理中的文本理解技術(shù)的研究及其應(yīng)用AbstractWiththedevelopmentofthesocialinformationtechn

2、ologypeople02esoeagertOcommunicatewiththecomputerbynaturallanguageNaturalLanguageUnderstandingisoneofthemostchallengingproblemsinthecomputerscience，anditisalsoonebreakthroughofthenewgenerationcomputerMoreandmorecountries

3、intheworldarepayingattentiontoitsresearch，andtextunderstandingbelongstotheapplicationsofthenaturallanguageunderstandingNowSOmanyorganizations，nomattertheyarebusinessenterprises，researchinstitutes，orgovernmentdepartments，

4、arefacingagreatdealofinformationtodealwithHowtohelptheuserslocateanddeliverinformationisonebigaimofinformationmanagementTheresearchobjectinthispaperistheChinesetextusedtOdescribethedetailsofacase，andwediscussedthemethods

5、fortextclassificationinnaturallanguageunderstanding，andrealizedthetextautomaticclassificationaccordingtOthecontentsInthispaperwedescribedthebasicprinciplesoftextclassification，mainalgorithmsandtechnologiesineachphaseofte

6、xtclassification，andthenmadesomecomparisonsWeimprovedthealgorithmoftextfeatureextraction，andalsointroducedthelatentsemanticanalysisusedtoinformationselection，whichcartdecreasethedimensionsoftextvectorspaceandimprovethecl

7、assificationperformanceAprimaryautomaticclassificationsystemforthelegaltextwascompletedinthispaperWdRenby：LiWen(ComputerSoftwareandTheory)Directedby：ProfW，angMinf1va—n—Keywords：NaturalLanguageUnderstanding，TextClassifica