This paper introduces the concept of spatial filters as an approach to increasing the relevance of the information retrieved by users of mobile information systems. This approach applies post-query filters to remove results that are deemed not relevant given some aspect of a mobile individual's spatial behaviour. The aim is to reduce the volume of results returned to users of mobile information systems, to avoid the need for manual filtering. These filters have been implemented in a location-based service (LBS) using georeferenced document collections from small numbers of content providers. Feedback from end-users in operational settings suggests that they found information that had been filtered in this way to be more relevant than unfiltered information. An opportunity for future research is to consider how these filters can be applied to the task of combining rankings based upon thematic and spatial criteria, for documents retrieved from unstructured or semi-structured collec...
David M. Mountain