Fuzzy dissimilarity-based classification for disaster initial assessment

Abstract

A correct initial assessment of disaster consequences is crucial for an adequate decision-making in disaster and emergency management. However, such an initial assessment needs to be correct, but not necessarily fully precise, and thus it can be associated with a fuzzy classification problem in which the set of classes presents a relevant structure. This paper proposes the consideration of a dissimilarity operator in order to introduce such a structure in the classifier's learning and reasoning procedures, leading to an improvement in the classifiers adaptation to the disaster management context features and decision making requirements.