How to use enriched browsing context to personalize web site access

Abstract

Using a browsing context is one of the keys to web site access personalization under particular constraints. With poor user information modeling, which is a common situation, a web site cannot be adapted to the current user. Assuming the current clickstream is the only known information about a web site user (no profile, no past sessions, no identification, no content analysis of viewed pages), we propose here a method to enrich the browsing context and enhance the current user model. In a batch mode, profile-based enriched navigation patterns are computed. In on-line mode, Navire, a personal agent and its matching rule engine continually re-adapts the browsing context with pre-calculated profiles. Based on the current up-to-date context, Navire personalizes the access to a web site.