WebWatcher: Knowledge Navigation in the World Wide Web

Many have noted the need for software to assist people in locating information on the World Wide Web. Although effective tools exist, they typically rely on brute-force scanning and indexing of Web pages for later keyword-based retrieval. Such tools ignore at least two sources of knowledge which might prove useful in navigation and retrieval: (1) the structure of the Web as a graph, and (2) the sequence in which human users search the web. We describe a learning apprentice system, called WebWatcher, which both performs the kind of indexing used by Web catalogers like Lycos, and attempts to exploit the two sources of knowledge listed above.