Metodología para mantener la interpretabilidad en el modelado utilizando sistemas borrosos

  1. Vélez Vélez, Miguel Ángel
Supervised by:
  1. Omar Sánchez Pérez Director

Defence university: Universidad de Huelva

Fecha de defensa: 02 December 2011

Committee:
  1. José Manuel Andújar Márquez Chair
  2. Manuel Jesús Maña López Secretary
  3. María Mercedes Pérez de la Parte Committee member
  4. Matilde Santos Peñas Committee member
  5. Emilio Jiménez Macías Committee member

Type: Thesis

Teseo: 325929 DIALNET lock_openArias Montano editor

Abstract

This PhD thesis is mainly aimed at improving interpretability in fuzzy adaptive models. Model interpretability can thus be improved with almost no loss of estimation accuracy. The new methodology proposed in this paper allows adjusting the membership functions of the fuzzy model. These functions can then be evaluated by an expert, thus leading to improved interpretability and semantics in their rules.