Similarity measurement in context

Abstract

Context plays a crucial role when measuring the similarity of two concepts. Nonetheless, the modelling of context has been mostly neglected in existing similarity measurement theories. In this paper, we explore the influence of context in existing similarity measurement approaches for the geospatial domain, focussing on whether and how these approaches account for it. Based on these observations, the processing of context during similarity measurement is analysed, and general implementation issues, especially ease of integration into existing reasoning systems and computability, are discussed. The results of the different analyses are then combined into a generic set of characteristics of context for similarity measurement, with regard to the geospatial domain. © Springer-Verlag Berlin Heidelberg 2007.