Hugh Osborne
Derek Bridge
School of Computing and Mathematics
Department of Computer Science
University of Huddersfield
University College, Cork

h.r.osborne@hud.ac.uk d.bridge@cs.ucc.ie

Abstract We motivate and present similarity metrics, the similarity measurement framework we have developed for use within Case-Based Reasoning Systems. In this framework similarities are values from any type on which a complete lattice is defined. This gives us a wide range of intuitive ways of measuring similarity and a large number of ways in which different metrics can be combined. The paper concludes by showing how the framework might be deployed in categorisation.