论文部分内容阅读
在名词词组译文的译文优化机制中,建立基于依存关系树的译文优化机制十分重要,如果缺乏类似机制,系统会输出与源语结构雷同的译文。显然,这种词与词一一对应的译文与目标语的文法结构不符,意义支离破碎,读者难以理解。因此,本文提出将依存关系树与计算机数据结构中相关概念结合,对输入名词词组进行句法剖析和标记;在此基础上与熟语料库中实例词组进行匹配,采用设定好的优化阈值,依据最相似熟语料词组对应的中文结构模式,输出经过译文优化机制的译文,以提高机器翻译译文的质量。
In the mechanism of translation optimization of noun phrase translation, it is very important to establish the translation optimization mechanism based on dependency tree. If there is no similar mechanism, the system will output the same translation as the source language structure. Obviously, this translation of the word and the word one by one does not match the grammatical structure of the target language, meaning is fragmented and difficult for the reader to understand. Therefore, this paper proposes to combine the dependency tree with the related concepts in the computer data structure to synthetically parse and mark the input noun phrase. Based on this, we match with the example phrase in the familiar corpus and adopt the set optimal threshold, Similar Chinese idioms are used to output the Chinese translation patterns that have been translated into the translation optimization mechanism to improve the quality of machine translation translations.