Abstract: It is very important to automatically acquire the head verb when analyzing a Chinese Sentence. Based on the theory of Hierarchical Network of Concepts, this paper put forward an approach to obtain the head verb candidates through the conceptual symbols of Chinese characters or words. Then the candidates were distinguished into different groups by some principles. One group, which is composed of the possible head verb candidates, will be sorted according their possibilities to be the head word. An experiment on 3121 Chinese sentences from real corpus showed that the 83% head verb candidates in the group were included in the head structure of the corresponding sentences.

Key words: Automatic Analysis of Chinese, Rule, Hierarchical Network of Concepts, Verb, the Head Verb

摘要： 语句核心动词的自动获取是以动词为中心的汉语语句分析中的重要组成部分。本文依据概念层次网络理论，在字词概念符号的基础上获取候选动词集合，根据动词的上下文语言环境对动词进行排除和分类处理，对可能作为语句核心动词的动词集合按照作语句核心结构的可能性大小排队，并验证其正确性。实验结果表明，在从真实语料切分得到的3121个语句中，经过排队处理后前三个动词作为语句核心动词的正确率达到了83%。错误分析表明进一步完善知识库及排除排队规则，还可以提高自动获取语句核心动词的正确率。

关键词: 汉语自动分析, 规则, 概念层次网络, 动词, 语句核心动词