A Location-Dependent Recommender System for the Web

Brunato, Mauro and Battiti, Roberto and Villani, Alessandro and Delai, Andrea (2002) A Location-Dependent Recommender System for the Web. UNSPECIFIED. (Unpublished)

Download (731Kb) | Preview


    A relevant piece of information in many context-aware applications for wireless and mobile users is the user's current location. Knowledge of the position, when combined with the user preferences, permits efficient service (or product) location, location-dependent alerting, and location-aware recommendation systems. We propose a recommendation system that is based on a standard web browser and where models determining the relevance of a given URL in a given region are derived in an automated and adaptive way through the collaboration of users of the system. With respect to existing location-dependent recommendation systems, the advantage of our proposal lies in the reduced effort required for system development and in the increased independence of the recommendation from the services (or products) owners. After an initial tuning phase, a specific URL will be recommended to a user in a given location in a way that considers where and how often it was accessed by the previous users. In detail, a new middleware layer, the "location broker", collects a historic database where user positions and links used in the past are analyzed to develop models relating resources to their spatial usage pattern and to calculate a preference metric when the current user is asking for recommendations. The focus of this work is on scalability issues. When the system is used in a wide area (possibly covering a sizable fraction of the entire web), the size of the database and the complexity of the models increase very rapidly. In an ubiquitous computing scenario where a multitude of wirelessly interconnected system surround a mobile user, the number of resources in an explored region may easily grow to contain thousands or millions of items. We describe a suitable data structure that permits scalability and analyze the empirical computational complexity both on a simulated scenario and in a real-world context in our province.

    Item Type: Departmental Technical Report
    Department or Research center: Information Engineering and Computer Science
    Subjects: Q Science > QA Mathematics > QA075 Electronic computers. Computer science
    Uncontrolled Keywords: Spatial databases, Context-aware computing, Recommendation systems, Collaborative filtering
    Report Number: DIT-02-093
    Repository staff approval on: 07 Jan 2003

    Actions (login required)

    View Item