Max-min Learning of Approximate Weight Matrices from Fuzzy Data
Abstract
In this article, we study the approximate solutions set of an inconsistent system of fuzzy relational equations . Using the norm, we compute by an explicit analytical formula the Chebyshev distance , where is the set of second members of the consistent systems defined with the same matrix . We study the set of Chebyshev approximations of the second member i.e., vectors such that , which is associated to the approximate solutions set in the following sense: an element of the set is a solution vector of a system where . As main results, we describe both the structure of the set and that of the set . We then introduce a paradigm for learning weight matrices that relates input and output data from training data. The learning error is expressed in terms of the norm. We compute by an explicit formula the minimal value of the learning error according to the training data. We give a method to construct weight matrices whose learning error is minimal, that we call approximate weight matrices. Finally, as an application of our results, we show how to learn approximately the rule parameters of a possibilistic rule-based system according to multiple training data.
Cite
@article{arxiv.2301.06141,
title = {Max-min Learning of Approximate Weight Matrices from Fuzzy Data},
author = {Ismaïl Baaj},
journal= {arXiv preprint arXiv:2301.06141},
year = {2023}
}