Soft computing

Method of fuzzy classifier design with self-tuning coevolutionary algorithms

R. B. Sergienko

M. F. Reshetnev Siberian State Aerospace University

Abstract: A new method of fuzzy classifier design that integrates Michigan and Pittsburgh approaches for fuzzy systems generating with evolutionary algorithms is presented. Coevolutionary approach for solving problem of genetic algorithm efficient parameters setting is used. Investigation of hybrid fuzzy classifier design for some practical classification problems is performed.

Keywords: fuzzy classifier, coevolutionary algorithm, classification, knowledge extraction.

Bibliographic databases:

•