基于句法分析的汉语词义消歧  

Chinese word sense disambiguation based on parsing analysis

作　　者：张春祥[1,2] 栾博[1] 高雪瑶[1] 卢志茂 

机构地区：[1]哈尔滨理工大学计算机科学与技术学院,哈尔滨150080 [2]哈尔滨理工大学软件学院,哈尔滨150080 [3]哈尔滨工程大学信息与通信工程学院,哈尔滨150001

出　　处：《计算机应用研究》2014年第1期40-42,共3页Application Research of Computers

基　　金：黑龙江省教育厅科学技术研究资助项目(12531106)

摘　　要：为了提高词义消歧的质量,对歧义词汇的上下文进行结构分析,提出了一种利用句法知识来指导消歧过程的方法。在歧义词汇上下文的句法树中,提取句法信息和词性信息作为消歧特征;同时,使用朴素贝叶斯模型作为消歧分类器。利用词义标注语料对分类器的参数进行优化,然后对测试数据中的歧义词汇进行消歧。实验结果表明,消歧的准确率有所提升,达到了66.7%。In order to improve the quality of word sense disambiguation, this paper analyzed the structure of the context including an ambiguous word and proposed a new method of word sense disambiguation based on parsing knowledge. It extracted parsing information and part of speech as disambiguation features. This paper used naive Bayesian model as word sense disambiguation classifier and trained its parameters on sense-annotated corpus. Then it applied the classifier to disambiguate ambiguous words in test data. Experimental results show that the accuracy rate of disambiguation is improved and arrives at 66. 7%.

关 键 词：词义消歧 句法信息 词性 消歧分类器 

分 类 号：TP391.2[自动化与计算机技术—计算机应用技术]