Typicality of Features in Fuzzy Relational Compositions

Abstract

Fuzzy relations and their compositions have the same crucial importance for the fuzzy mathematics that is provided to mathematics by relations and their composition. The topic, since it attracted many scholars and influenced many areas in fuzzy modeling, has been extended on distinct directions including the recent one on the incorporation of excluding features. This article brings a mathematically similar yet semantically opposite extension, in particular, the concept of typical features. We show the appropriateness of such a new concept and investigate some of its properties. Furthermore, we discuss how the concept of typical features incorporated in fuzzy relational compositions may bring significant improvement of results of some applications. This fact is demonstrated on a real example of biological species classification.