Recommendations Based on Region and Spatial Profiles

Fuelled by the quantity of available online spatial data that continues to grow, the requirement for filtering spatial content to match mobile users’ context becomes increasingly important. This paper introduces a flexible algorithm to derive users’ preferences in a mobile and distributed system. Such preferences are implicitly computed from users’ virtual and physical interactions with spatial features. Using this concept, region profiles for specific spatial contexts can be generated and used to recommend content to those visiting that region. Our approach provides a set of profiles (personal and region-based) which are combined to adapt the presentation of a given service to suit users’ immediate needs and interests. A proposed college campus navigation assistant illustrates the benefits of such an unobtrusive recommender system.