Abstract: Existing results of auxiliary word are difficult to use in the automatic annotation of natural language processing. Based on the auxiliary words knowledge base, rule-based method is used in automatic annotation of auxiliary words usage. Contrast to the results of test, it shows that refining, extension and adjusting the matching order of the rules can promote the precision and recall effectively. It is also benefit for improve the quality of Chinese Corpus, deepen the processing depth, and reduce the artificial work.

Key words: auxiliary words, knowledge base, usage, rule, automatic annotation

摘要： 目前已有的助词研究成果很难直接应用于自然语言处理的机器识别。在现代汉语词典、规则库、语料库“三位一体”的助词知识库基础上，采用基于规则的方法进行了现代汉语常用助词用法的自动识别。对比规则优化前后的实验结果证明，对用法的规则进行细化、扩充和调序可以有效地提高助词用法识别的准确率和召回率，减轻人工标注的工作量，提高大规模语料库的质量。

关键词: 助词, 知识库, 用法, 规则, 自动识别