A recursive algorithm for fuzzy Min-Max networks

In the present paper, a new algorithm to train Min-Max neural models is proposed. It is based on the ARC technique, which overcomes some undesired properties of the original Simpson's algorithm. In particular, training results do not depend on pattern presentation order and hyperbox expansion is not limited by a fixed maximum size, so that it is possible to have different covering resolutions. ARC generates the optimal Min-Mar network by a succession of hyperbox cuts. The generalization capability of ARC technique depends mostly on the adopted cutting strategy. A new recursive cutting procedure allows ARC technique to yield a better performance. Some real data benchmarks are considered for illustration.