摘要
词义排歧在自然语言处理领域占有重要地位。词义排歧的精确率依赖于排歧知识的完备性。但是目前使用的基于词典的和基于语料库的词义排歧方法来获取排歧知识的效果都不令人满意。文中借助《知网》,以义原同现频率矩阵作为排歧知识,在其基础上设计并实现了一个基于义原同现频率的汉语词义排歧系统,大大地提高词义排歧的精确率。
Word sense disambiguation has always been important in natural language processing, The precision rate of word sense disambiguation depends on the completeness of disambiguation knowledge. Most methods currently which are baesd on dictionary and on corpora are not satisfiable. This paper introduces an automatic system of Chinese word sense disambiguation which is based on HOWNET and takes primitive co - occurrence data matrix as the disambiguation knowledge. The restdt of experimentation proves that the the precision rate of word sense disambignation is higher.
出处
《计算机技术与发展》
2006年第5期184-185,188,共3页
Computer Technology and Development
关键词
自然语言处理
词义排歧
义原
natural Larguage processing
word sense disambiguation
primitive
