摘要
在计算机科学和语言学中,针对动词语义层面上的分类问题,研究者们提出了不同的分类方法,但这些分类方法无一例外地都存在着分类不全面等分类学中经常遇到的问题.一个动词表示一个事件,该文以获取事件相关的常识知识为出发点,以动词性语素为分类依据对常见的现代汉语动词进行语义分类,此分类方法分类标准清晰、不丢失语义信息,并且可实现自动分类,该文重点以"自移"类动词为例来介绍我们的分类方法.此外,该文用描述逻辑对动词性语素及语素之间的分类关系进行形式化表示,动词性语素的形式化表示是动词形式化表示的基础.根据该事件语义分类结构,可以有效地进行事件属性常识知识和事件关系常识知识的获取.
Abstract
Since a verb denotes an event,we classify common modern Chinese verbs at a semantic level based on verb morphemes for the purpose of acquiring commonsense knowledge of events.In contrast to the existing linguistic verb classification taxonomy,the categorization criteria of this method is clear:preserving the semantic information con-tained in this verb after a verb is classified.This paper utilizes the class of"self-motion events"as an example to in-troduce our categorization method.We use description logics to formalize the verb morphemes and relationships be-tween these verb morphemes,which is essential to formalization of verbs.According to this event classification sys-tem,we can effectively acquire commonsense knowledge related to event attributes and event relationships.
基金项目
国家重点研究与发展计划(2017YFC1700300)
国家重点研究与发展计划(2017YFB1002300)
国家自然科学基金(61702234)