Automated knowledge acquisition for a fuzzy classification problem

Abstract

Genetic algorithms and neural networks are useful as automated knowledge acquisition tools for Fuzzy Systems. This paper describes the application of these techniques to a well known classification problem, namely the iris species classification problem. The performance of the resulting fuzzy systems exceed that reported for those derived using alternative methods. Preliminary work indicates that the use of genetic algorithms is the more flexible as it allows the simultaneous acquisition of fuzzy set parameters and fuzzy rules.