Browsing Support by Extraction and Ordering of Partial Documents in Web Sites


@In recent years, the growth of the World Wide Web has been rapidly expanding. In particular,
the number of pages at each site continue to increase and the structure of the sites is becoming
more complex. Therefore, it is not always easy to retrieve information even if the users find
the site which they want to visit. Since the search engine do not understand the actual meaning
of the content by browsing the pages only, there are many cases where users would not be
satisfied by visiting the pages. The results acquired from search engines is a list of page
titles in which keywords are contained i.e. since each page is treated as an independent
document in such reference algorithm, structures of the Web sites are disregarded. The
structures of the Web sites means the combination of pages and the linkage of anchors of each
page. In other words, it is a hypertext organization itself. The structure of the Web site must
reflect the author's intention. The most important factor of structure is navigation among
pages. That is, it is a fundamental design methodology where anchors are put in the pages and
in which the navigation manner is prepared.

Generally, users have to view many pages by browsing one by one in a Web site. In fact, pages
with a great deal of descriptions are contained. In such a situation, it is not easy to
understand the structure and the contents of the site and to specify portions which fit with
the user's requirements.

In this research, in order to solve these problems, we propose an automatic navigation method
following the user's requirement in Web sites. This method detects the essential pages in the
site and generates the order list of the pages to present results by linkage information of the
pages. We have designed and implemented a prototype system based on our proposed method.
In the current version, the functionality of automatic navigation is implemented by Flash movie
contents.