基于统计的句法分析方法研究  

Research on method of syntax parsing based on statistical approach

作　　者：王继曾[1] 任浩征[1] 罗恒[1] 刘宽[1] 

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《计算机工程与设计》2006年第12期2207-2210,共4页Computer Engineering and Design

摘　　要：在概述统计句法分析主要特征的基础上,对目前主流分析方法进行了深入分析,提出了实用化的基于统计的句法分析器必须引入结构信息、上下文信息、词汇信息以及语义信息才能提高性能和可靠性的观点。总结了基于统计的汉语句法分析方法当前的研究现状,最后探讨了基于统计的汉语句法分析方法的不足之处和发展趋势。Based on the feature of statistical syntax parsing, the prevalent methods are analyzed deeply and the viewpoint is obtained that analytic measures, such as structure, context, lexical entries, semantics and etc, are introduced in order to improve the performance and reliability of the practical statistical syntax parsers. Then the current research of Chinese statistical parsing is summarized. Finally, the disadvantages of Chinese statistical parsing and its development trend are discussed.

关 键 词：句法分析 统计 模型 语料库 汉语 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]