NaviMoz: mining navigational patterns in portal catalogs

Portal Catalogs is a popular means of searching for information on the Web. They provide querying and browsing capabilities on data organized in a hierarchy, on a category/subcategory basis. This paper presents mining techniques on user navigational patterns in the hierarchies of portal catalogs. Specifically, we study and implement navigation retrieval methods and clustering tasks based on navigational patterns. The above mining tasks are quite useful for portal administrators, since they can be used to observe users' behavior, extract personal preferences and re-organize the structure of the portal to satisfy better user needs and navigational habits. These mining tasks have been implemented in the NaviMoz, a prototype system for mining navigational patterns in portal catalogs