Author(s)

A. Platis, P. Tselios G. Vouros

Abstract

This paper, in a greater extend than previous work, proposes a method for identifying website pages that may be of high interest to website visitors but are not, somehow, directly accessible by them, and for providing navigational support. Towards this account we model website visitors activity in a site using a Markov chain and provide a method for measuring pages' transitions attractabilities on a time interval of site usage. In this way, we have an indicator for page usage depending on the information interest of the page, the probability of a user to follow a link to that page and the probability of a user to be in that page. The paper provides examples from a test case and discusses the utilization of attractabilities for re-organising a we

Keywords