The behavior of an e-customer varies greatly. For companies, the understanding of this behavior is a major challenge to better meet the expectations of the customer, try to convince him to buy their products, to build loyalty and develop their capital.

Why does a consumer choose one product over another? What makes him prefer one brand to another? Why does he visit the same product more often? Why is the purchase made during the lunch break? What are the factors that influence his choice and his purchases? To know more about customer behavior, further analysis is needed. The data to be analyzed are available. These are all the data left by the user during his visit to an e-commerce site. A fine analysis of the customer's behavior consists of understanding his journey on the site.

The aim of this thesis is to use mathematical models, graph models and algorithms to model and analyze log files of client traces.