COFRE: a fuzzy rule coevolutionary approach for multiclass classification problems

This paper presents a technique for solving multiclass classification problems using a revolutionary approach. There are m populations, where m is the number of classes. Individuals from different classes work together to define a classifier, while individuals from the same population compete among them to define the best rule for a class. Finally, the best classifier is selected according to its performance. Experiments are conducted on different publicly available data sets.