Research for WWW Information Service in Digital Archives of Japanese Literature

为了在因特网上检索日本文学领域的研究资料--翻译文本、象形文字图像数据等日本文学的数字档案，开发了WWW信息系统。首先，为了识别是否存在数字档案，使用书名、人名、地名等该领域的专业术语，分析了存在这些数字档案的已知网站的页面描述。然后我们提取了描述的典型模式。在此基础上，我们开发了jlda_filter（Japanese Literary Digital Archive_filter），这是一个特定于领域的过滤器，用于提取足以判断是数字存档网站的描述。利用该过滤器，我们开发和评价了一个WWW信息系统，该系统包括两个部分：一是利用本领域的专业术语词典对URL进行筛选，二是利用jlda_filter判断网站是否有预期的存档。一是通过对各档案的描述，如哪些是原书、哪些是不同版本、谁翻译的、谁是所有者等，进一步分析，给出各档案的来源和特征，作为元数据。二是利用我们的日本古典文学书籍的位置信息数据库，同时实现了在因特网上检索原书和电子书的位置信息的功能。

We developed the WWW Information system to search the digital archives of Japanese Literature, such as translated text and hieroglyphic image data, which are research materials in this field on Internet.At the first to recognize if there is digital archives, we analyzed the page description of known sites which have such digital archives by using technical terms of this domain, such as book name, person name, and place name. Then we extracted the typical pattern of descriptions. Based on this analysis, we developed jlda_filter ( Japanese Literary Digital Archive_filter ) which is the domain specific filter to extract the description enough to judge that is the digital archiving site. By using this filter, we developed and evaluated WWW Information system consists of two parts; the first is to narrow down URL by using technical term dictionaries in this domain and the second is to judge if the site has expected archives by using jlda_filter.In the next, we have two plans. The one is to give origin and feature of each archive as metadata by analyzing descriptions more, for example, which is the original book and different editions, who translated, and who is the owner, etc... The other is to give the function to retrieve both locations of original book and its electric copy on Internet by using our Location Information DB of classical Japanese Literary books simultaneously.