Abstract: The automatic recognition of Chinese modality’s usage is an important task for building Chinese modality knowledge base.This paper focuses on the problem of the recognition of Chinese modality’s usage for multi-corpus.It modifies the modality’s usage dictionary and the modality’s usage rules base from the real usage of Chinese modality of multi-corpus.The experiments show that the recognition precision rate of modality’s usage is higher than before and the applicability of Chinese modality knowledge base is enhanced.

Key words: natural language processing, functional word knowledge base, modality, rules, corpus

摘要： 语气词用法的自动识别是现代汉语语气词知识库的核心问题。使用规则方法研究了语气词用法在多种语料库中的识别问题，从多种语料库中的语气词实际用法入手，修改和完善了语气词用法词典和语气词用法规则库。实验数据表明，经过修改和完善，语气词用法在各语料库中的识别准确率有所提高，增强了语气词知识库的适用性。

关键词: 自然语言处理, 虚词知识库, 语气词, 规则, 语料库