数据库受限汉语自然语言查询的分词研究与实现

The research and implementation of word segmentation of database natural language query based on restricted Chinese

对数据库受限汉语自然语言查询语句进行分词处理.分词算法分为两个部分,第一部分对最大匹配法进行改进,改进的核心思想是体现整句长词优先的原则,改进后的算法能够减少切分歧义;第二部分根据实例数据库的查询需要处理姓名和不稳定的属性值两类未登录词,未登录词的识别对后续句子的理解起着至关重要的作用.

This paper describes the word segmentation of database natural language query based on restricted Chinese. The word segmentation algorithm is made up of two parts. The first part improves the maximum matching segmentation algorithm that fully embodies the principle of priority of long word on a whole sentence. The improved algorithm can decrease the ambiguity of segmentation. The second part processes two sort of unlisted words that are name and unstable property value according to the requirement of instance database. The recognition of unlisted words plays an important role in the following understanding to sentence.