多语种翻译词汇的在线自动抽取
Multilingual Translation Extraction from the Internet
-
摘要: 越来越多网页以多种语言的形式在互联网上传播 ,从中抽取多语种翻译词汇具有重要的研究价值 针对网页的特点 ,提出了一种新的多语种翻译词汇的在线自动抽取方法 该方法通过对双语网页中超链接信息相似度的计算 ,获取多语种翻译词汇 ,相似性越高 ,对应的词条互为翻译对的可能性越大 通过对中英、德英、法英 3类双语网页的抽取 ,结果证明它具有较高的准确率 ,是一种高效的与语言无关的多语种词汇对抽取方法Abstract: Multilingual translations play a vital role in natural language processing Based on multilingual Web pages on the Internet, a novel method to extract translations from hyperlinks is presented The more similar the both hyperlinks are, the greater the translation possibility of both terms is The experiments on English Chinese/German/French Web pages prove that the method is correct and also language independent The average precision of extraction is about 90%
下载: