Developing a Knowledge Network of URLs

1 Introduction WWW is a huge database. Discovering a new knowledge from such a database is an important theme. Search engines are commonly used and clustering search results are proposed in 2]. However, those information are used temporarily and only kept in user's bookmark. The bookmark does not represent the whole structure of user's knowledge. Forming a new knowledge structure is much harder and valuable. This paper describes the system KN(Knowledge Network) which visualizes the linkage structure of a user's personal knowledge of URLs. It is an extension the system in 3]. A personal knowledge of URL-links is drawn as a directed graph called as a Web graph. Some softwares display the static structure of a particular web site. The linkage information on WWW are also studied (e.g..1, 4, 5]). Our database keeps not only general linkage information, but also personal knowledge. The system supports personal and progressive linkage database. Our system stores background knowledges as Web graphs. Given a background knowledge and some keywords, the system gains new URLs from a search engine with the keywords and places them on the given background knowledge. A new URL is placed around an old URL when they are directly linked. In this way, a user can develop his knowledge and can keep the new knowledge as another background knowledge.