Extracting hyponymy relation between Chinese terms based on term types'commonality and sequential patterns
ICIC Express Letters(2009)
摘要
This paper studies the problem of the automatic acquisition of the hyponymy(is-a) relation in sentences and proposes a new method for it. The previouslydeveloped pattern-based approaches, to our knowledge, require that terms havingthe target relation occur syntactically near one another, i.e., in the samesentence. In this paper, we treat the task of identifying hyponymy relation astwo separate problems and propose a method which could identify hyponymyrelation among terms that don't co-occur in a sentence. Besides, we introducethe Sequential Patterns to decide terms' type, which enables us to representpatterns consisting of distant words in sentences. Experimental results showthat our method is promising and could recognize more relation instances. ICIC International © 2009.
更多查看译文
关键词
Hyponymy relation,Relation extraction,Sequential pattern,Term types'commonality
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络