A Comparison of Selectional Preference Models for Automatic Verb Classification

A Comparison of Selectional Preference Models for Automatic Verb Classification

ID:40254546

大小:190.19 KB

页数:12页

时间:2019-07-29

A Comparison of Selectional Preference Models for Automatic Verb Classification _第1页
A Comparison of Selectional Preference Models for Automatic Verb Classification _第2页
A Comparison of Selectional Preference Models for Automatic Verb Classification _第3页
A Comparison of Selectional Preference Models for Automatic Verb Classification _第4页
A Comparison of Selectional Preference Models for Automatic Verb Classification _第5页
资源描述:

《A Comparison of Selectional Preference Models for Automatic Verb Classification 》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、AComparisonofSelectionalPreferenceModelsforAutomaticVerbClassificationWillRobertsandMarkusEggInstitutfürAnglistikundAmerikanistik,HumboldtUniversity10099Berlin,Germany{will.roberts,markus.egg}@anglistik.hu-berlin.deAbstractverbargumentpreferences.SPsarerepresentedbyfeaturesca

2、pturinglexicalinformationabouttheWepresentacomparisonofdifferentselec-headsofargumentstotheverbs;werestrictourtionalpreferencemodelsandevaluatethemfocusheretonouns.onanautomaticverbclassificationtaskinWeoperationaliseaselectionalpreferencemodelGerman.Wefindthatallthemodelsweas

3、afunctionwhichmapssuchanargumentheadcompareareeffectiveforverbclustering;toaconceptlabel.Wesubmitthattheprimarythebest-performingmodelusessyntacticcharacteristicofsuchamodelisitsgranularity.Ininformationtoinducenounsclassesfromourbaselinecondition,allnounsaremappedtotheunlab

4、elleddatainanunsupervisedman-samelabel;thiseffectivelycapturesnoinformationner.Averysimplemodelbasedonlexicalaboutaverb’sSPs(i.e.,weclusterverbsusingsub-preferencesisalsofoundtoperformwell.categorisationinformationonly).Ontheotherex-treme,eachnounisitsownconceptlabel;weterm1

5、Introductionthisconditionlexicalpreferences(LP).BetweenthebaselineandLPlieaspectrumofmodels,inSelectionalpreferences(KatzandFodor,1963;whichmultipleconceptsaredistinguished,andWilks,1975;Resnik,1993)arethetendencyforeachconceptlabelcanrepresentmultiplenouns.awordtosemantical

6、lyselectorconstrainwhichOurmainhypothesisisthatverbclusteringwillotherwordsmayappearinadirectsyntacticre-workbestusingamodelofsuchintermediategran-lationwithit.Selectionalpreferences(SPs)haveularity.ThisfollowstheintuitionthatverbswouldbeenaperennialknowledgesourceforNLPtask

7、sseemtoselectforclassesofnouns;forinstance,suchaswordsensedisambiguation(Resnik,1997;wesupposethatessen‘eat’wouldtendtopreferasStevensonandWilks,2001;McCarthyandCar-adirectobjectanounfromtheabstractconceptEs-roll,2003)andsemanticrolelabelling(Erk,2007);sen(‘food’).Weassumeth

8、attheseconceptscanbeandrecognisingselectionalviolationsisthoughtexpressedin

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。