X-FCS: a fuzzy classifier system using accuracy based fitness - first results

A fuzzy classifier system called X-FCS is proposed which employs accuracy-based fitness. Each classifier (rule) maintains an estimate of the payoff obtained when it fires and the accuracy of this prediction is used as the basis for selection under action of the genetic algorithm. The motivation behind this approach is that such a classifier system will be capable of generating compact, high-performance rule sets which are simultaneously general, accurate and coadapted.