摘要
在分析目前分词方法的基础上提出了一种通过建立多元信息库、采用改进型的粗分算法以找出所有可能存在歧义的句子、借助于人工干预建立错误切分歧异词库等,实现汉语歧异切分的方法,通过修改、插入多元信息库中的信息量,进一步设计了一个具有自适应能力的歧义切分方法,并通过实验证明该方法能够有效改进汉语分词中错误歧义切分的结果.
On the basis of the analysis of the existing algorithms of Chinese word segmentation, the article puts forward to realize Chinese word ambiguous segmentation by establishing mulit-gram library and improving the rough segmentation algorithm in order to find all sentences which have ambiguous word segmentation and establishing false segmentation ambiguous word library in virtual of manual interference. Moreover, the article designs a self-adaptive Chinese ambiguous word segmentation method based on modifying and inserting the values in the multi-gram library. It proves that the new method is able to improve effects in false ambiguous word segmentation.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第8期1597-1600,共4页
Journal of Chinese Computer Systems
基金
江苏省高技术研究项目(BG2005020)资助
江苏省自然基金项目(BK2003030)资助
江苏省教育厅自然(04KKB320134)资助
关键词
多元信息库
歧义切分
自适应
multi-gram library
ambiguous segmentation
self-adaptive
