Dynamic system identification using recurrent neural network with multi-valued connection weight

Abstract

This paper introduces a new concept of the connection weight to the standard recurrent neural networks - Elman and Jordan networks. The architecture of the modified networks is the same as that of the original recurrent neural networks. However, in the modified networks the weight of each connection is multi-valued, depending on the value of the input data involved. The backpropagation learning algorithm is also modified to suit the proposed concept. The modified networks have been benchmarked against their original counterparts. The results on eleven benchmark problems are very encouraging.