摘要
中文是一种话题结构的语言,其表达方式比较灵活,但句法结构不如英文严谨,导致了事件中论元与触发词的关系较松散。现有的论元抽取方法多数是基于浅层语义的句法结构特征,从而造成了论元抽取性能低下。为了解决这个问题,提出了基于语义的中文事件论元抽取方法。该方法利用角色、实体和触发词的语义,弥补了论元抽取中单纯采用句法特征的缺陷。在ACE2005中文语料上的测试结果表明,该方法与基准系统相比具有更高的性能。
Chinese is a topic-based language,and the expression of Chinese sentences is very flexible,which leads to the loose connection between the arguments and the trigger in an event.Because the syntactic features based on shallow semantic are widely used in the previous studies,it leads to the low performance of argument extraction.To solve this issue,this paper put forward a novel argument extraction approach based on multiple semantics.This approach makes use of the semantic of roles,entities and trigger to be a supplement of syntactic-based approaches.Experimental results on the ACE 2005 Chinese corpus show that our approach outperforms the baseline significantly.
出处
《计算机科学》
CSCD
北大核心
2015年第2期237-240,共4页
Computer Science
关键词
论元抽取
角色语义
论元语义
Argument extraction
Semantics of role
Semantics of argument