Visualization and collaborative filtering for web mining tasks

The paper proposes several possible tasks for analyzing a web site HTML contents and corresponding Web-Log file using approaches from text visualization and collaborative filtering areas. For this purpose, we used only a simplistic version of the Web-Log file information: a set of Web-Page user accesses in the form of unordered pairs of user-and page-identifications. We developed a recommendation system that enables sharing information about the visits of other users to the same Web page and a system for clustering of the users based on the similarity of their behavior. Additionally, we also cluster the Web pages based on their content and based on the user visits.