Equivalence between neural networks and fuzzy systems
A.E. Gaweda, Jacek M. Żurada · 2002
Demonstrates that a single-hidden layer feedforward neural network is equivalent to a fuzzy inference system with relational rule antecedents. The method establishes a link between networks weights and fuzzy system parameters and defines the upper bound on the number of fuzzy rules required to represent the network. An application example illustrates the proposed idea.