in [2] discussed a neural network with regularized negative correlation learning RNCL methodology, as the neural network has high sensitivity for its initial condition; therefore, this method is combined with the output of different networks which gives better performance over error reduction.