基于依存树的越南语新闻事件元素抽取

Extracting Vietnamese news event elements based on dependency tree

为加强对越南政治、经济和文化等方面新闻事件的了解,提出一种基于依存树的越南语新闻事件元素抽取方法。分析越南语的语法特点,发现越南语最主要的特点是定语后置,其它和中文的语法结构类似,通过直接映射中文句法结构得到越南语依存树;在此基础上通过定义规则,在依存树中找到相应的句法结构,抽取句子的主语、宾语和状语。实验结果表明,该方法可以快速地定位到越南语句子的句法成分,有效地抽取出越南语新闻事件元素。

To enhance the understanding of the Vietnamese political,economic and cultural aspects of news events,a method extracting Vietnamese news event element based on dependency tree was proposed.According to Vietnamese grammar characteristics,the facts that the main feature of Vietnamese is attributive post position and other grammatical structures are similar to Chinese were found,so Vietnamese dependency tree was got by directly mapping Chinese sentence structure.On this basis,by defining rules,the syntactic structure in dependency tree was found,thereby extracting the subject,object and adverbial of a sentence.Experimental results show that this method can quickly locate the syntactic constituents of Vietnamese sentence and effectively extract the Vietnamese news event element.