Abstract: This paper focuses on the automatic recognition of Chinese modality usage. It builds the modality usage dictionary and the modality usage rules base from the real usage of Chinese modality. It introduces the concept of semantic field when building the modality usage rules base. It does experiments to automatically recognize Chinese modality usage by using the modality usage rules base. The corpus is the People’s Daily in January 1998, and the correct rate of label is 78.433%.

Key words: natural language processing, functional word knowledge base, recognition of modality usage, semantic field

摘要： 研究现代汉语语气词用法的自动识别问题，从语气词的实际用法入手，构建语气词用法词典和语气词用法规则库，利用语气场构建语气词用法规则库。采用1998年1月的《人民日报》分词与词性标注语料，利用语气词用法规则库进行语气词的自动识别。实验结果证明，语气词的识别准确率为78.433%。

关键词: 自然语言处理, 虚词知识库, 语气词用法识别, 语义场

CLC Number: 

• TP391