Modelling the behavior of individual websurfers
Abstract Book of the XXIII IUPAP International Conference on Statistical Physics, Genova, Italy, (9-13 July 2007)Abstract
We perform a large scale analysis of the behavior of individual users as they navigate through a World Wide Web domain. The logs of the webserver responsible for serving Emory University's web content under the emory.edu domain and all subdomains over a period of $294$ consecutive days is analyzed. The resulting bipartite graph contained $3,179,671$ IP addresses connected to $2,562,398$ unique URLs through $53,582,121$ individual and timestamped page requests. Limited information on the referer URLs is also available. The topology and the dynamics of the network evolution is analyzed and both numerical and analytical modeling strategies are proposed. The resulting properties of the simulated network are compared with the original empirical results.