基于译文的英汉双语句子自动对齐

Translation－Based Automatic Alignment of English and Chinese Parallel Corpora

双语语料库的自动对齐已成为机器翻译研究中的一个重要研究课题.目前的句子对齐方法有基于长度的方法和基于词汇的方法,该文先分析了基于长度的方法,然后提出了基于译文的方法：通过使用一部翻译较完整的词典作为桥梁,把英汉句子之间的对应关系连接起来.根据英语文本中的单词,在词典中找到其对应的译文,并以译文到汉语句子中去匹配,根据评价函数和动态规划算法找到对齐句对.实验结果证明这种对齐方法消除了基于长度做法中错误蔓延的情况,它大大地提高了对齐的精度,其效果是令人满意的.

Parallel corpora alignment is a important research issue in MT.The current sentence alignment methods include length－based and lexicon－based methods.This paper analyses length－based method and proposes translation－based method：With a dictionary as the bridge,corresponding translation of the words in English text are found in the dictionary and then matched with the translation of Chinese text.The alignment sentences are determined according to an evaluating function and the dynamic programming algorithm.The experimental result shows that this new method greatly improves the precision of sentence alignment．