Fuzzy Classifiers Tuning Through an Adaptive Memetic Algorithm

This paper presents a methodological approach for tuning the fuzzy rules of a fuzzy classifier using an adaptive memetic algorithm. The approach is validated over two benchmark problems in terms of classification error and computational effort. In addition, it compares the performance of memetic, genetic and adaptive memetic algorithms over the benchmark problems. These results show a favorable trend towards the tuning of the classifiers through the adaptive memetic perspective.