Using machine learning to enhance software tools for Internet information management

This paper discusses the issues involved in the appli-cation of machine learning techniques to the manage-ment of Internet-based information. We present a gen-eral architecture, and describe how this has been in-stantiated in several different applications. The first three of these systems provide assistance to a user sort-ing incoming mail, reading USENET news or iden-tifying World-Wide Web pages of interest; the final system constructs a personalised on-line newspaper, assembled from documents gathered by a Web robot. A number of machine learning techniques have been used in the construction of these systems; some com-parative results are presented