双语法律文本的自动对齐

Automatic Alignment of Bilingual Law Text

基于实例的机器翻译系统需要双语句对的支持。为大量获取双语句对,则需要以篇章对齐的双语文本为输入,实现句子的自动对齐。通过分析汉英双语法律文本的特征,提出了法律文本对齐假设。首先识别出法规源文和译文中的结构标识和句子,然后在句子一级对齐法律文本。该方法在150篇汉英法律文本语料上,取得了80.98%的对齐准确率。

Example-based machine translation system needs the support of bilingual sentences.In order to acquire enough bilingual sentences,it is necessary to align sentences from chapter-aligned bilingual text automatically.After analyzing the structure characteristic of Chinese-English law text,the alignment hypothesis of law text is presented.The method recognizes the structure markers and sentences in the law source and translation text,then aligns them according to the alignment hypothesis.The experiment corpora ...