Representing generalized fuzzy automata in recurrent neural networks
M. Doostfatemeh, Stefan C. Kremer · 2004
In this paper we present a new architecture for the representation of general fuzzy automata (GFA). It is based on second-order recurrent neural networks (2ORNN). The architecture implements the functions F/sub 1/ and F/sub 2/, used in GFA, into the structure of 2ORNN. The performance of this representation method is compared with a previous method for embedding fuzzy automata, and it is shown that GFA are more efficiently representable in 2ORNN.