2010 | OriginalPaper | Buchkapitel
A Multi-objective Neuro-evolutionary Algorithm to Obtain Interpretable Fuzzy Models
verfasst von : Gracia Sánchez, Fernando Jiménez, José F. Sánchez, José M. Alcaraz
Erschienen in: Current Topics in Artificial Intelligence
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In this paper, a multi-objective constrained optimization model is proposed to improve interpretability of TSK fuzzy models. This approach allows a linguistic approximation of the fuzzy models. A multi-objective evolutionary algorithm is implemented with three different selection and generational replacements schemata (Niched Preselection, NSGA-II and ENORA) to generate fuzzy models in the proposed optimization context. The results clearly show a real ability and effectiveness of the proposed approach to find accurate and interpretable TSK fuzzy models. These schemata have been compared in terms of accuracy, interpretability and compactness by using three test problems studied in literature. Statistical tests have also been used with optimality and diversity multi-objective metrics to compare the schemata.