Authors:
Ku Muhammad Naim Ku Khalif
and
Alexander Gegov
Affiliation:
University of Portsmouth, United Kingdom
Keyword(s):
Interval Type-2 Fuzzy Sets, Uncertainty, Defuzzification, Vectorial Centroid, Machine Learning, Bayesian Logistic Regression, Human Intuition.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Fuzzy Systems
;
Soft Computing
;
Type-2 Fuzzy Logic
Abstract:
It is necessary to represent the probabilities of fuzzy events based on a Bayesian knowledge. Inspired by
such real applications, in this research study, the theoretical foundations of Vectorial Centroid of interval
type-2 fuzzy sets with Bayesian logistic regression is introduced. This includes official models, elementary
operations, basic properties and advanced application. The Vectorial Centroid method for interval type-2
fuzzy set takes a broad view by exampled labelled by a classical Vectorial Centroid defuzzification method
for type-1 fuzzy sets. Rather than using type-1 fuzzy sets for implementing fuzzy events, type-2 fuzzy sets
are recommended based on the involvement of uncertainty quantity. It also highlights the incorporation of
fuzzy sets with Bayesian logistic regression allows the use of fuzzy attributes by considering the need of
human intuition in data analysis. It is worth adding here that this proposed methodology then applied for
BUPA liver-disorder dataset and va
lidated theoretically and empirically.
(More)