一种全切分与统计结合的分词系统  

Chinese Segmentation System Combining Omni-Segmentation with Statistic

作　　者：王荔[1] 宋胜利[1] 冯佳[1] 陈平[1] 

机构地区：[1]西安电子科技大学软件工程研究所,陕西西安710071

出　　处：《微电子学与计算机》2009年第5期68-70,共3页Microelectronics & Computer

基　　金：国防"十一五"预研项目(513060601)

摘　　要：歧义消解是中文分词的主要问题之一.提出了一种全切分与统计结合的分词算法,构造出基于统计词典的有向无环词图,利用动态规划算法得出最佳的分词路径.实验证明,系统有效地提高了歧义切分的准确性及分词速度.Ambiguity resolution is one of the main problems in Chinese word segmentation.This paper presents a Chinese segmentation system combining omni-segmentation with statistic.A directed acyclic graph based on statistical dictionary is first constructed,and then the best segmentation path is obtained by dynamic programming algorithm.The experiments show that both the accuracy of ambiguous segmentation and the speed of the segmentation are improved effectively in this system.

关 键 词：中文分词 全切分 统计分词 歧义消解 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]