Uncertainty-aware similarity measures – properties and construction method

Abstract

A concept of uncertainty-aware similarity measure is being defined and discussed. The aim of the paper is to support the opinion that both definition and construction of such measure should take into account the epistemic nature of compared incomplete (uncertain) information. This approach, however, generates new challenges resulted from the computational complexity of the problem. We define a set of properties to be satisfied by uncertainty-aware similarity measure and we propose a new technique of constructing such measures for Interval-Valued Fuzzy Sets.

On the Search for a Measure to Compare Interval-Valued Fuzzy Sets

“…However, this perspective is not the usual one. To the best of our knowledge, the literature where we can find this approach is rather limited [30][31][32][33]. These papers can be considered as the starting point of our research [34].…”