Harmonisation of soft logical inference rules in distributed decision systems

Abstract

It is considered a problem of harmonisation of diagnostic rules used in a distributed set of diagnostic centres, the rules being based on a k-most-similar-cases approach. Harmonisation is reached due to an exchange of diagnostic cases among the reference sets stored in the centres. The method is based on general concepts of similarity, semi-similarity and structural compatibility measures used to evaluation of adequacy of records in remote data files to the requirements connected with supporting decision making in a given, local diagnostic centre. The procedure of local reference set extension by diagnostic cases selection and acquisition is described. © Springer-Verlag Berlin Heidelberg 2006.