摘要
利用语义、语法等语言知识,建立一种分层句法分析统计模型,并进行句法分析实验。研究结果表明:该模型具有规则和统计相结合的特点,且在层次分析的不同阶段,根据不同的语法、语义、语用特性采用不同的方法和不同的统计模型;该模型结合分词、词性标注进行句法分析,是一个词汇化的句法分析模型,可同时考虑多个语义依存关系;采用该模型,精确率和召回率分别为87.23%和86.15%,其综合指标F与头驱动句法分析模型的相比提高了5.25%。
By incorporating linguistic features such as semantic dependency and syntactic relations, a novel statistical parsing model was proposed, and experiments were conducted for the refined statistical parser. The results show that the mode not only takes advantage of linguistic features such as semantic dependency or syntactic relations, but also considers context such as adjoining words. The model can take advantage of a few semantic dependencies at the same time. It is a parser based on lexicalized model. It achieves 87.23% precision and 86.15% recall using the model, and comprehensive index F is improved by 5.25% compared with that using the head-driven oarsing model.
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2012年第3期986-991,共6页
Journal of Central South University:Science and Technology
基金
国家自然科学基金资助项目(60763001)
江西省自然科学基金资助项目(2009GZS0027
2010GZS0072)
关键词
自然语言处理
词聚类
中心词驱动
句法分析统计模型
natural language processing
word clustering
head-driven parsing model
statistical parsing model
