A Method for Analyzing Bookmark to Support Information Retrieval

摘要

It is expected to propose a technique to obtain only useful information for each user. Because,it is quite difficult for users to obtain their desired information among a large amount of search results returnee! by a Web search engine. Therefore, it is expected to propose a technique for obtaining only useful information for each user. In this paper, we propose an assistant system for user to help users to search Web pages using both collaborative filtering and filtering exploiting the structure of web bookmark. In our study, we regard each user's bookmark as a collection of useful knowledge, so that, we consider not only bookmarked Web pages but also its structure as useful knowledge. Based on these ideas, first we develop a collaborative filtering system that recommends Web pages bookmarked by other users using keywords included in a user's browsing Web page. And then our system carefully select the recommended Web page using filtering system based on the strucuture of bookmark and propose the selected Web pages as final recommended Web pages.