Pattern Detection with Improved Preprocessing in Web Log

The past fifteen years are characterized by an exponential growth of the Web both in the number of Web sites available and in the number of their users. This growth generated huge quantities of data related to the user’s interaction with the Web sites, recorded in Web log files. Moreover, the Web sites owners expressed the need to better understand their visitors in order to better serve them. The Web Use Mining (WUM) is a rather recent research field and it corresponds to the process of knowledge discovery from databases (KDD) applied to the Web usage data. It comprises three main stages: the preprocessing of raw data, the discovery of schemas and the analysis (or interpretation) of results. A WUM process extracts behavioral patterns from the Web usage data and, if available, from the Web site information (structure and content) and on the Web site users (user profiles). In this thesis, we bring two significant contributions for a Web Use Mining process. We propose a customized application specific methodology for preprocessing the Web logs and a modified frequent pattern tree for the discovery of patterns efficiently.