Ambiguity Resolution for Vt-N Structures in Chinese

Ambiguity Resolution for Vt-N Structures in Chinese

ID:40354421

大小:436.77 KB

页数:10页

时间:2019-07-31

Ambiguity Resolution for Vt-N Structures in Chinese  _第1页
Ambiguity Resolution for Vt-N Structures in Chinese  _第2页
Ambiguity Resolution for Vt-N Structures in Chinese  _第3页
Ambiguity Resolution for Vt-N Structures in Chinese  _第4页
Ambiguity Resolution for Vt-N Structures in Chinese  _第5页
资源描述:

《Ambiguity Resolution for Vt-N Structures in Chinese 》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、AmbiguityResolutionforVt-NStructuresinChinese1,221Yu-MingHsiehJasonS.ChangKeh-JiannChen1InstituteofInformationScience,AcademiaSinica,Taiwan2DepartmentofComputerScience,NationalTsing-HuaUniversity,Taiwanmorris@iis.sinica.edu.tw,jason.jschang@gmail.comkchen@iis.sinica.edu.twmallyaVPstructure.Ho

2、wever,ChineseverbsAbstractcanalsomodifynounswithoutmorphologicalinflection,e.g.,養殖/farming池/pond.Conse-Thesyntacticambiguityofatransitivequently,parsingVt-Nstructuresisdifficultbe-verb(Vt)followedbyanoun(N)hascauseitishardtoresolvesuchambiguitieswith-longbeenaprobleminChineseparsing.outpriork

3、nowledge.ThefollowingaresomeInthispaper,weproposeaclassifiertotypicalexamplesofvariousVt-Nstructures:resolvetheambiguityofVt-Nstructures.Thedesignoftheclassifierisbasedon1)threeimportantguidelines,namely,解決/solve問題/problemVPadoptinglinguisticallymotivatedfeatures,解決/solving方案/methodNPusinga

4、llavailableresources,andeasyin-解決/solve人類/mankind(問題/problem)Nonetegrationintoaparsingmodel.Thelin-Tofindthemosteffectivedisambiguationfea-guisticallymotivatedfeaturesincludetures,weneedmoreinformationabouttheVt-Nsemanticrelations,context,andmorpho-NPconstructionandthesemanticrelationslogic

5、alstructures;andtheavailablere-betweenVtandN.StatisticaldatafromtheSini-sourcesaretreebank,thesaurus,affixda-caTreebank(Chenetal.,2003)indicatesthattabase,andlargecorpora.Wealsopro-58%ofVt-Nstructuresareverbphrases,16%posetwolearningapproachesthatresolvearenounphrases,and26%donothaveanyde-the

6、problemofdatasparsenessbyauto-pendentrelations.ItisobviousthatthesemanticparsingandextractingrelativerelationsbetweenaVt-Nstructureanditscon-knowledgefromlarge-scaleunlabeledtextinformationareveryimportantfordifferen-data.Ourexperimentresultsshowthattiatingbetweendependentrelations.Althoughth

7、eVt-Nclassifieroutperformsthecur-theverb-argumentrelationofVPstructuresisrentPCFGparser.Furthermore,itcanbewellunderstood,itisnotclearwhatkindofse-easilyandeffectivelyintegratedintothemanticrelationsresultinNPstructures.InthePCFGparserandgene

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。