A system to restructure hypertext networks into valid user models

We have implemented an experimental system that automatically restructures hypertext networks according to their users’ browsing behavior. The system applies link weights to the hyperlinks in the networks and updates these link weights according to three learning rules. The learning rules are based on how often a particular hyperlink is being traversed and operate on strictly local information of link traversals. Changes in network structure are fed back to users by dynamic link ordering according to descending link weight. The system has been shown to be able to structure random hypertext networks into valid representations of their users’ browsing preferences in two WWW experiments and a simulation using a mathematical model of user navigation.