基于统计的翻译等价词对抽取研究

Translation Equivalent Pairs Extraction Based on Statistical Measures

对利用汉英双语语料进行翻译等价词对抽取进行了探索 .对不同的词对关联度量方法进行了对比 ,考察了词性信息、多词翻译单元识别对翻译等价词对抽取结果所可能产生的改进以及负面影响 .

This paper examines how the translation equivalent pairs could be extracted from parallel corpus. Four different word association measurements are compared with each other. And this paper also proposes to use part of speech information for performance improvement and integrate a multi-word unit identification component to alleviate the problem of partial correct result. Some negative aspects of these proposals are also examined in this paper.