Aid of navigation through recommendation

In this paper, we describe an approach for automatically finding the prototypic navigation behavior of Web users. User access logs are examined in order to extract the most significant user navigation access pattern. Such approach gives us an efficient way to better understand the way users are acting, and leads us to improve the structure of Web sites for improving navigation. Another interesting application deals with recommender systems where the system can suggest to users links that could interest them. A main issue of navigation pattern discovering deals with the way to decide when two traces are identical or not, even when they present some little difference. In this paper we describe the overall characteristics of our similarity function, and we will introduce some of our results.