基于弱化语法规则的机器翻译方法研究

在机器翻译的方法中，基于规则的翻译方法和基于实例的翻译方法都有各自的优缺点。结合两者的优点，提出了一种基于弱化语法规则的机器翻译方法，该方法在大量分析句子语法特征的基础上以及利用语言专家的知识构建语法规则库，在利用语法识别出句子后，把句子的各个组成部分标记出来，然后利用语义块进行匹配推理翻译出句子的各个组成部分。最后根据句法把句子的各个译出的组成部分装配起来形成译文。实验表明，该方法达到了较好的翻译效果，并且具有较大的发展潜力。

An Approach of Machine Translation Based on Weaken Syntax

Among methods of machine translation,both rule-based machine translation and example-based machine translation have their own advantages and disadvantages.In this paper,a machine translation method based on weaken syntax is proposed which makes use of the advantages of both the rule-based and example-based methods.This method builds a syntax database by analyzing lots of sentences to extract rules and summarizing language expert's knowledge.It identifies the modules of the sentence which we want to translate,and translates these modules into target language.At the end it outputs the translation by assembling translated modules.The experiment shows that the proposed algorithm has better performance.