首页 | 本学科首页   官方微博 | 高级检索  
     

基于语言学的依存分析结果动宾关系补全研究
引用本文:张丹.基于语言学的依存分析结果动宾关系补全研究[J].计算机应用研究,2018,35(4).
作者姓名:张丹
作者单位:沈阳航空航天大学人机智能中心
摘    要:格语法理论中一般动词带有必须格,而汉语中右侧带必须格的动词映射到依存句法中该动词一般带有宾语。基于该理论识别依存句法分析结果中缺失动宾关系依存弧的现象,针对该类错误,提出了一种基于知识约束的算法,寻找缺失依存弧以改进依存句法分析。实验结果表明,利用语言学知识能有效地识别依存句法分析结果中缺失动宾关系依存弧的动词,识别准确率达到93.80%,改进后的系统UAS值提升了0.21%,动宾关系UAS值提升了2.14%,说明利用语言学知识能够改善依存句法分析效果。

关 键 词:动宾关系    语言学知识  依存句法分析
收稿时间:2016/12/15 0:00:00
修稿时间:2018/2/23 0:00:00

Research on Completion of Verb-object Relation in Dependency Parsing Based on Linguistic Knowledge
Affiliation:Human Machine Intelligence Research Center,Shenyang Aerospace University,
Abstract:In the case grammer, the verb has the necessary case, and the verb with necessary case on the right in Chinese has object in Chinese dependcency syntax. This paper identifies missed dependency arc of verb-object relation in dependency parsing analysis result based on this theory, and solves the problem through a method based on knowledge constraints to impove dependency parsing. The experiment results show that, using the linguistic knowledge can identify the verbs with missed dependency arc of verb-object relation and the recognition accuracy is almost up to 93.80%, the UAS of improved system increased by 0.21%, and the UAS of verb-object relation increased by 2.14%. It shows that the use of linguistic knowledge can improve the dependency parsing performance effectively.
Keywords:verb-object relation  linguistic knowledge  dependency parsing
点击此处可从《计算机应用研究》浏览原始摘要信息
点击此处可从《计算机应用研究》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号