Fuzzy multi-layer perceptron for binary pattern recognition

This paper presents a modified fuzzy multilayer perceptron model which uses, as a basis, the implementation developed in Pal and Mitra (1992). The proposed model uses different parameters in the calculation of a fuzzy desired output vector to Pal and Mitra as well as adding an ambiguity parameter to the updating equation in order to take into account the degree of ambiguity of an input pattern during the learning phase. Handwritten character recognition is discussed in particular.