Modeling and Application Based on Diagonal Recurrent Neural Network

A simple recurrent neural network named as diagonal recurrent neural network was studied. To overcome the slow convergence of BP algorithm, the recursive prediction error (RPE) algorithm was proposed, which can train both the weight and the bias. A given model was identified by using diagonal recurrent neural network trained with RPE algorithm, and the model of a phosphating temperature control system was established. Both simulation and experiment demonstrate the effectiveness of the proposed algorithm.