基于词典和语素的交集型歧义消除模型  

Elimination of Chinese overlapping ambiguity based on dictionary and morpheme

作　　者：李春雨[1] 王勇 

机构地区：[1]浙江机电职业技术学院,浙江杭州310053 [2]杭州茂亨自控仪表有限公司,浙江杭州310053

出　　处：《微型机与应用》2013年第4期12-14,18,共4页Microcomputer & Its Applications

摘　　要：提出了一种消除中文分词中交集型歧义的模型。首先通过正向最大匹配法和逆向最大匹配法对中文文本信息进行分词,然后使用不单独成词语素表对分词结果进行分析对比消歧,得到符合汉语语境的结果。整个过程分为歧义识别、歧义分析、歧义消除三个阶段。实验结果表明,该模型可以有效降低由交集型歧义引起的中文文本切分错误率。This paper puts forward an overlapping ambiguity elimination model of Chinese segmentation. Firstly, the model segments the Chinese text corpus by maximum matching method and reverse maximum matching method, and then analyzes the segmentation to eliminate the ambiguity based on non-single morpheme list to obtain correct result which fits the Chinese context. The totally processing includes three sections as following： ambiguity detection, ambiguity analysis and ambiguity elimination. The experiment data indicates that the model referred in the paper reduces Chinese segmentation mistakes caused by overlapping ambiguity effectively.

关 键 词：自然语言处理 分词 交集型歧义 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]