Abstract
This research is a part of developing a search system providing a suitable retrieval among categorized useful information for higher education. This paper focuses on automatic categorization, that is, automatic extraction of information from the Web sites using keywords contained in the anchortext and page link structures. The information in the laboratory pages in Tsukuba University can be extracted with high precision by gathering not only the starting page linked from the anchortext but also other pages linked from the starting page within the same URL directory. This method using anchortexts and page link structures would be applied to the automatic extraction of web pages of which text and context analysis are not applicable.
