Comparison of Fuzzy ARTMAP and MLP Neural Networks for Hand-Written Character Recognition

Abstract

Fuzzy ARTMAP is one of the recently proposed neural network paradigm where the fuzzy logic is incorporated. In this paper, we compare the Fuzzy ART MAP neural network and the well-known backpropagation based Multi-layer perceptron (MLP), in the context of hand-written character recognition problem. The results presented in this paper shows that the FuzzyARTMAP out-performs its counterpart, both in learning convergence and recognition accuracy.