基于DOP技术的目标语生成机制

提出在面向数据的英汉机译系统中，一种以面向数据的语言分析技术作为基本框架的目标语生成机制。该机制通过对源语语句的句法分析树进行线性化操作，生成目标语译文。其中包括从源语语句句法分析树的所有片段组合形式中选择一个适合生成操作的生成片段组合形式、对生成片段组合形式中的所有片段进行线性化操作以及对所有片段已经线性化的生成片段组合形式进行线性操作，从而获取最终的目标语译文。为论证方法有效性，基于包含1，000个语句的真实英语语料构建知识源，并采用包含100个语句的真实英语语料作为测试集。实验表明，目标语译文质量比较令人满意，可成功地实现英汉机译。

IMPLEMENTING TARGET LANGUAGE GENERATION BASED ON DOP TECHNIQUE

This paper presents a kind of target language generation mechanism in Data Oriented English Chinese Machine Translation System. This mechanism applies DOP technique which is used in language analysis traditionally into target language generation equally. Through linearizing source language analysis result syntax tree, the final translation in target language is generated. This process includes selecting a generation fragment combination form which is appropriate to generation operation from all the fragment combination forms of the syntax tree of the source language sentence, linearizing all the fragments in the generation fragment combination form and the generation fragment combination form itself and acquiring the final translation in target language. To prove the efficiency of the proposed method, the knowledge source is constructed based on the real world English corpus which involves 1,000 English sentences, and the other real world English corpus which includes 100 English sentences is used as the test set. The experiment result shows that the quality of translation in target language is satisfactory and the English Chinese machine translation process can be implemented successfully.