基于匹配模板的术语自动翻译方法

Method for automatic translation of terms based on template matching

本文针对于专利术语的特点,建立了一个基于属于模板的术语自动翻译系统。系统中结合了翻译模型和相似度打分机制,实验结果表明该方法能够解决术语翻译中的固定术语翻译问题和基于中心词序调整问题,提高了统计机器防疫的质量。同时术语模板中引用了单一的统计翻译中增加了语言学知识,为翻译提供了一个有效手段。

this paper according to the characteristics of patent term, set up an automatic translation system belongs to the term based on template. The system combines a translation model and a similarity scoring mechanism, experimental results show that this method can solve the problem of fixed tern1 translation in translation and adjustment based on the center word order, improve the quality of statistical machine epidemic prevention. At the same time the term template cited increased linguistic knowledge in a single statistical translation, provides an effective means for translation.