Mining Web logs for a personalized recommender system

As the Web rapidly grows, however, the number of matching pages increases at a tremendous rate when users use the search engine for finding some information. It is not easy for a user to retrieve the exact information he/she requires. In particular, browsing a Web set is an expensive operation, both in time and cognitive effort. Recommender systems have then become valuable resources for users seeking intelligent ways to search through the enormous volume of information available to them. In this paper we propose a new framework based on Web logs mining for building a personalized recommender system. At the core of personalization is the task of building a profile of the user. We have developed an approach that user's information learned from user's Web logs data to construct accurate comprehensive individual profiles. One part of this profile contains facts about a user, and the other part contains rules describing that user's behavior. We use Web usage mining to derive the behavioral rules from the data.