Modelling and predicting a Web page accesses using Markov processes

The significance of modelling and measuring various attributes of the Web in part or as a whole is undeniable. Although Web related metrics have become increasingly sophisticated, few employ models to explain their measurements. In this paper, we discuss metrics for usage characterization. We considered the application of patterns in browsing behavior of users for predicting access to Web documents. We proposed two models based on Markov processes for addressing our specification of the access prediction problem. The first adapts a stochastic model for library circulations, i.e., Morse's model in the context of accessing Web documents. The second approach can be used for determining access probabilities of Web pages within a site by modelling the browsing process as an ergodic Markov chain.