Generally speaking, usage-based Web personalization systems [10] involve 3 phases: data preparation and transformation, pattern discovery, and recommendation. Of these, the latter is a real-time component, while the other two phases are performed offline. The pattern discovery phase may include the discovery of association rules, sequential navigational patterns, clustering of users or sessions, and clustering of pageviews or products. The recommendation engine considers the active user session in conjunction with the discovered patterns to provide personalized content. The personalized content can take the form of recommended links or products, or targeted advertisements tailored to the user's perceived preferences as determined by the matching usage patterns. In this paper, we focus on a specific instance of this general framework in which the recomendations are produced based on matching the current user session against patterns discovered through association rule mining on user transaction data. First, we briefly discuss the data preparation and pattern discovery phases and then we focus on the details of our recommendation engine.
Bamshad Mobasher (mobasher@cs.depaul.edu)