Abstract:

The models based on the Markov property have been widely used for modelling the users’ navigational behaviour in a web site. In this paper we propose an algorithm that computes the preferential navigational path. The model can be used for web path personalization on a website in order to meet the  characteristics and preferences of the users. Markov models have been  widely used for modeling the users’ navigational behavior in a web site. The models based on the Markov property are based on the transition probabilities between web pages, as recorded in web log files. Web usage mining represents the set of techniques used for the study of the users’ behavior during navigation on the Internet. A common problem modeled by the Markov chains is computing the navigational paths used by the users during their navigation on a web site [6,10].  Though, Markov models do not  handle the case when a path is not included in the training data or is included in low frequency and cannot provide good estimates of the corresponding transition probabilities.