The BINGO! Focused Crawler: From Bookmarks to Archetypes

bove (i.e., viewingthe query terms as an initial training document).According to [CBD99a] the key components of afocused crawler are a document classifier to test whethera visited document fits into one of the specified topics ofinterest, and a distiller to identify the best URLs for thecrawl frontier (i.e., those hyperlinks in already visiteddocuments that, when traversed, promise the best resultsin the continuation of the crawl). Obviously the distillershould be aware of the...