摘要： 介绍了一个对汉语语料进行切分标注粗加工的系统SegPos。该系统采用前缀码分词算法，用二元语法模型进行词性标注，并利用概率统计、规则、歧义数据库、部分句法分析等多种方法的结合排除分词和标注中产生的歧义。

关键词: 切分, 标注, 语料库, 自然语言理解

Abstract: This paper presents the authors' recent work towards development of SegPos, a Chinese corpus segmentation & part of speech tagging system. SegPos does segmentation with prefix -method and does tagging with 2-meta syntax model. And it uses many methods such as statistics, rules, ambiguity table and part syntax analysis to decrease the ambiguity appeared during the process of segmentation and tagging.

中图分类号: 

• TP391.12