Premium
Properties and comparison of andness‐characterized aggregators
Author(s) -
Dujmović Jozo,
Torra Vicenç
Publication year - 2021
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.22346
Subject(s) - news aggregator , fuzzy logic , computer science , theoretical computer science , mathematics , implementation , artificial intelligence , programming language , operating system
Logic aggregators are all aggregators characterized by andness/orness. In this paper we present necessary properties of logic aggregators and use them to compare major implementations of andness‐characterized aggregators: means, t ‐norms/conorms, ordered weighted average family, fuzzy integrals, and graded conjunction/disjunction. Our goal is to provide methodology for justifiable selection of the most suitable aggregator for various information fusion problems. When aggregating arguments from [0, 1], such arguments are regularly interpreted as degrees of truth or degrees of fuzzy membership. In such cases we deal with logic aggregators that must have appropriate logic properties. Our analysis identifies 10 necessary properties that are critical for decision support applications and must be satisfied by logic aggregators. Then, we evaluate and compare five families of logic aggregators that offer different levels of support to desired logic properties.