Paper Info
Title
A new approach to uncertainty description through accomplishment membership functions
Abstract
We propose an alternative constructor for fuzzy membership functions based on statistics.Proposed methodology is consistent with possibility theory and attains certainty limits calculation and boundaries.A link is made between ellipsoidal rules, certainty levels, and data statistics. Human linguistic reasoning and statement logic are able to solve uncertain propositions. Similar capabilities are expected to be found on intelligent systems so they are provided with some sort of artificial logic evaluation. Many approaches to uncertainty measurement have been developed before, mainly referring to probability or possibility theories. Some conceptual restrictions are imposed by forcing a distribution function to be conceptually consistent. In this work, conditions imposed to possibility theory are relaxed and the theoretical perspective is oriented to degrees of accomplishment. Conceptual implications and their relation to numerical calculations with respect to a specific class of membership functions are presented. Relation to possibility theory and certainty measurement are discussed to show logical consistency, together with a synthetic numerical example which helps to elaborate conclusions about data dispersion and its relation to the accomplishment proposal.
Year
DOI
Venue
2015
10.1016/j.eswa.2015.05.021
Expert Systems with Applications
Keywords
Field
DocType
Fuzzy sets,Membership functions,Data analysis,Ellipsoidal rules,Uncertainty measurements,Possibility theory
Mathematical economics,Certainty,Intelligent decision support system,Computer science,sort,Fuzzy logic,Measurement uncertainty,Fuzzy set,Possibility theory,Artificial intelligence,Membership function,Machine learning
Journal
Volume
Issue
ISSN
42
21
0957-4174
Citations 
PageRank 
References 
1
0.43
26
Authors
3
Name
Order
Citations
PageRank
Luis Ibarra111.11
Pedro Ponce22418.14
Arturo Molina364269.86