An evolving neuro-fuzzy recurrent network

In this research, we propose an evolving neuro-fuzzy recurrent network (ENFRN). The network is capable to perceive the change in the actual system and adapt (self organize) itself to the new situation. The network generates a new hidden neuron if the smallest distance between the new data and all the existing hidden neurons (the winner neuron) is more than a given radius. We propose a new pruning algorithm based on the density. Density is the number of times each hidden neuron is used. If the value of the smallest density (the looser neuron) is smaller to a specified umbral, this neuron is pruned. We use a modified least square algorithm to train the parameters of the network. Structure and parameters learning are updated at the same time. The major contribution of this research is: we present the stability of the algorithm of the evolving neuro-fuzzy reccurrent network proposed. Two simulations give the effectiveness of the suggested algorithm.