首页> 外文会议>International conference on computational linguistics;COLING 2010 >An Optimal and Portable Parsing Method for Romanian, French, and German Large Dictionaries
【24h】

An Optimal and Portable Parsing Method for Romanian, French, and German Large Dictionaries

机译:罗马尼亚语,法语和德语大型词典的最佳便携式分析方法

获取原文

摘要

This paper presents a cross cross-linguistic analysis of the largest dictionaries cu cur- rrently existing for Romanian, French, and rently German, and a new, robust and portable method for Dictionary Entry Parsing (DEP), based on Se Segmentation mentation- Cohesion Cohesion-Dependency (SCD) configur configura- ations. The SCD configurations are a tions. ap- pplied successively on each dictionary e plied en- ntry to identify its lexicographic se try segments ments (the first SCD configuration), to extract its sense tree (the second conf configuration), guration), and to parse its atomic sense definitions (the third one). Using pr previous results on vious DLR (The Romanian Thesaurus ?new format), the present paper adapts and a ap- pplies the SCD plies SCD-based technology to other four large and complex thesauri: DAR (The Romanian Thesaurus ?old format), TLF (Le Tr閟or de la Langue Fran?Fran鏰ise), aise), DWB (Deutsches W鰎terbuch ?GRIMM), and GWB (G鰐he G鰐he- W鰎terbuch). This experiment is ill illu- ustrated on significantly large parsed e strated en- ntries of these thesauri, and proved the fo tries fol- llowing features: lowing (1) the SCD SCD-based m me- ethod is a co thod completely pletely formal grammar grammar- free approach for dictionary parsing, with pproach efficient (weeks weeks-time adaptable) mode model- ling through sense hierarchies and parsing ing portability for a new dictionary. (2) SCD SCD- configurations separate and run seque sequen- ntially and independently the processes of tially lexicographic seg segment re ment recognition, sense ognition, tree extraction, and atomic definition parsing. (3) The whole DEP process with SCD SCD-configurations is optimal optimal. (4). SCD SCD- configurations, through sense marker classes and their dependency hype hyper- rgraphs, offer an unique instrument of le graphs, lex- xicon construction comparison, sense co icon con- ncept design and DEP standardiz cept standardization. tion.
机译:本文介绍了针对目前罗马尼亚语,法语和少量德语的最大词典的跨语言分析,以及一种基于Se Segmentationmentation- Cohesion的新的,健壮且可移植的字典条目解析(DEP)方法。凝聚依赖性(SCD)配置配置。 SCD配置是一项。依次应用到每个字典上,尝试识别其词典顺序段(第一个SCD配置),提取其意义树(第二个conf配置),guration)并解析其原子意义定义(第三个)。利用先前在以前的DLR(罗马尼亚词库新格式)上得到的结果,本论文将SCD层基于SCD的技术改编并应用到其他四个大型而复杂的词库中:DAR(罗马尼亚词库旧格式), TLF(Le Tr閟or de la Langue Fran?Fran鏰ise,aise),DWB(Deutsches W鰎terbuch? GRIMM)和GWB(G鳄he G鳄he- W鰎terbuch)。该实验不适用于这些叙词表的大型解析条目,并证明了以下功能:降低(1)基于SCD SCD的方法是完全完全正式的方法语法语法分析的无语法方法,通过感知层次结构和解析新字典的可移植性,以pproach高效(数周至数周的时间自适应)模式建模。 (2)SCD SCD-配置顺序地依次分离和运行,并且独立地执行字典字典段识别,感觉识别,树提取和原子定义解析的过程。 (3)具有SCD SCD配置的整个DEP过程是最优的。 (4)。 SCD SCD配置通过感知标记类及其依赖关系的超图,提供了一种独特的工具,包括图表,词典结构比较,感知图标概念设计和DEP标准化概念标准化。 tion。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号