Web user modeling can be done in many different ways. Below will be described some of them and will be provided the links to the resources.
One of the way of modeling web user behavior is to use Markov chains. This theory allows us to find the probability of clicking available links if we know the transition matrix. Elements of this matrix are the probabilities of moving from page i to page j
Clustering is the assignment of group of objects into subgroups (clusters) based on some metric. This can be also used to cluster users into some subgroups and then we can compare new user path with the cluster which is the most close to this user. More detailed description of clustering is provided at [1]
Vector space models use similarity to compare user data. Each user can be represented by visited pages, by some keywords from visited pages or some other data. With this model the user is represented as the vector in the vector space. Some perl code and examples for vector space modeling can be found at [2]
Soft computing like neural networks, genetic algorithm, swarm intelligence and many other heuristic or AI methods also can be used for web user modeling. [3] provides taxonomy for personalization of web-based systems by CI models. Collaborative filtering is a method of making automatic predictions (filtering) about the interests of a user by collecting taste information from many users (collaborating). [4]
Different types and many algorithms are described on Wikipedia site. Thus there are many ways to do web user modeling. This post provides some ideas and links to resources that can be used as the starting point.
References
1.Cluster Analysis, From Wikipedia, the free encyclopedia http://en.wikipedia.org/wiki/Cluster_analysis
2. http://www.lwebzem.com/cgi-bin/res/user_modeling_1.cgi Website User Modeling with Perl
3. RECENT ADVANCES in COMPUTER ENGINEERING and APPLICATIONS Personalization of Web-Based Systems based on Computational Intelligence Modeling TRICIA RAMBHAROSE, ALEXANDER NIKOV
4 Collaborative filtering, From Wikipedia, the free encyclopedia http://en.wikipedia.org/wiki/Collaborative_filtering