资源描述：
《Reuse of a Proper Noun Recognition System in Commercial and Operational NLP Applications》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。
1、ReuseofaProperNounRecognitionSysteminCommercialandOperationalNLPApplicationsChinatsuAoneandJohnMaloneySRAInternational4300FairLakesCourtFairfax,VA22033aonec@sra.com,maloneyj~,~sra.cornAbstractspecificallythechoiceofhigh-recallorhigh-precisionstrategies.Butfirst,wedis
2、cusstherelevantfeaturesSRA'sproprietaryproduct,NameTagTM,ofNameTag.whichprovidesfastandaccuratenamerecognition,hasbeenreusedinmanyap-plicationsinrecentandongoingefforts,in-2DescriptionofNameTagcludingmultilingualinformationretrievalandbrowsing,textclustering,andassis
3、-NameTagisamultilingualnamerecognitionsystem.tancetomanualtextindexing.ThispaperItfindsanddisambiguatesintextsthenamesofpeo-reportsonSRA'sexperienceinembeddingple,organizations,andplaces,aswellastimeandnamerecognitioninthesethreespecificap-numericexpressionswithveryh
4、ighaccuracy.Thedesignofthesystemmakespossiblethedynamicplications,andthemutualimpactsthatoc-cur,bothonthealgorithmiclevelandinrecognitionofnames:NameTagdoesnotrelyontherolethatnamerecognitionplaysinuserlonglistsofknownnames.Instead,NameTagmakesuseofaflexiblepatternsp
5、ecificationlanguagetointeractionwithasystem.Inthecourseofthis,wetouchuponvariousinteractionsidentifynovelnamesthathavenotbeenencounteredbetweenpropernamerecognitionandma-previously.Inaddition,NameTagcanrecognizeandlinkvariantsofnamesinthesamedocumentauto-chinetransla
6、tion(MT),aswellastheroleofaccuratenamerecognitioninimprovingmatically.Forinstance,itcanlink"IBM"to"Inter-theperformanceofwordsegmentational-nationalBusinessMachines"and"PresidentClin-ton"to"BillClinton."gorithmsneededforlanguageswhosewrit-ingsystemsdonotsegmentwords.
7、NameTagincorporatesalanguage-independentC-t-+pattern-matchingenginealongwiththelanguage-specificlexicons,patterns,andotherre-1Introductionsourcesnecessaryforeachlanguage.Inaddition,FastandaccuratenamerecognitionproductsaretheJapanese,Chinese,andThaiversionsintegrateo
8、nlynowcomingontothemarket.SRA'spropri-wordsegmenterstodealwiththeorthographicchal-etaryproduct,NameTag,hasbeenreusedinmanylengesoft
