On the equivalence of neural networks and fuzzy expert systems

It is proven that any continuous, layered, feedforward neural net can be approximated to any degree of accuracy by a (discrete) fuzzy expert system, and that any continuous, discrete, fuzzy expert system with one block of rules may be approximated to any degree of accuracy by a three layered, feedforward neural net. The second result may be generalized to multiple blocks of rules by considering total (discrete) input and total (discrete) output from the fuzzy expert system. It is concluded that fuzzy expert systems and neural nets can both approximate functions (mappings, systems).<<ETX>>