摘要
确定性依存句法分析(又称为基于动作的依存句法分析或基于转换的依存句法分析),由于采用了贪婪的搜索策略,它是一类比较有效率的依存句法分析方法。在一种主流的确定性依存句法分析方法—Yamada-Matsumoto算法的基础上,通过改进其转换模型,提出了一种基于子树的确定性依存句法分析方法。在新的方法中,引入了子树的结构,因而能够更加灵活地确定词与词间的依存关系。最后,通过在中文宾州树库上的实验,证明了这个新方法的有效性。
Deterministic dependency parsing（also known as action-based dependency parsing or transition-based dependency parsing） is an efficient type of dependency parsing methods,because the greedy strategies are used to search optimal dependency structures.On the basis of a mainstream deterministic dependency parsing method,Yamada-Matsumoto algorithm,we propose a state-of-the-art deterministic dependency parser,the sub-tree based deterministic dependency parsing method by improving its transition model.In our new method,optimal dependency between words can be searched in a more flexible way because of the newly introduced sub-tree structure.The effectiveness of the method is demonstrated with the experiments on Chinese Penn Treebank.
出处
《计算机应用与软件》
CSCD
2011年第2期268-270,共3页
Computer Applications and Software
