摘要： 提出一种扩展的关联规则法用于抽取中文非分类关系，在利用普通关联规则抽取出非分类关系概念对后，通过语言学规则抽取相应的非分类关系名称。该方法克服了普通关联规则法无法得出具体非分类关系名称的缺点，能够确定非分类关系的定义域和值域。实验结果表明仅使用扩展的关联规则法进行非分类关系抽取即可完成任务，无须对概念对与“谓语”动词之间的紧密程度进行验证。

关键词: 本体学习, 非分类关系, 关联规则

Abstract: This paper introduces an extended association rule which learns Chinese non-taxonomic relation names with linguistics rule after extracting concept pairs with general association rule. This method overcomes the disadvantage of not knowing the name of non-taxonomic relation names with general association rule, confirms the domain and range. Experimental results show that the extended association rule can reach the aim of task well and no need more compact validation between concept pair and its related predicate verb.

Key words: ontology learning, non-taxonomic relation, association rule

中图分类号: 

• TP311