基于web的日英词典自动获取技术的研究

双语词典资源对于自然语言处理研究的巨大价值已经得到越来越多的认可，尤其是在机器翻译、跨语言信息检索等方面。传统人工构建词典存在规模小、时效性差、语种不够灵活等问题，而且耗费大量的人力财力。提出了一种基于web的获取日英释义的新方法。该方法采用双向融合计算技术，首先对日语单词进行启发式搜索，在返回的网页摘要中应用基于距离和频率的方法挖掘出其候选英语单词；在第二阶段对候选英语单词进行二次搜索，并对两个阶段的结果进行融合计算来确定日英释义。

A web-based automatic acquisition method for Japanese-English dictionary

It has been increasingly recognized that bilingual dictionary has a great value for natural language processing research, especially in machine translation, cross-language information retrieval, etc. Dictionary manually built by traditional methods is faced with such problems as small scale, poor timeliness, inflexibil- ity, and high investigative cost. This paper presents a new web-based method of obtaining Japanese-English interpretation. The method, using two-way integrated computing technology, firstly takes the Japanese words for heuristic search and applies the method based on distance and frequency to dig out their English candidates in the returned summary pages; secondly it takes the English candidates for the second round search and performs integrated calculation of the two-stage results to determine the Japanese-English inter- pretation.