Case-based reasoning using dominance-based decision rules

Abstract

Case-based Reasoning (CBR) is a process of inferring conclusions related to a new situation by the analysis of similar cases known from the past experience. We propose to adopt in this process the Dominance-based Rough Set Approach (DRSA), that is able to handle monotonicity relationships of the type "the more similar is object y to object x with respect to the considered features, the closer is y to x in terms of the membership to a given fuzzy set X". At the level of marginal similarity concerning single features, we consider this similarity in ordinal terms only. The marginal similarities are aggregated within decision rules underlying the general monotonicity property of comprehensive closeness of objects with respect to their marginal similarities. © 2011 Springer-Verlag.