An analysis of performance of feed forward neural network: using back propagation learning algorithm

In some practical applications Neural Network (NN), a fast response to external events within extremely short period is required. However, using back propagation (BP) based on gradient descent optimization method obviously not satisfy many applications because of serious problems with BP are slow convergence speed of learning and containment low minima. Over the years, many improvements and modifications to the learning algorithm BP have been reported. In this research, we modified existing BP learning algorithm with adaptive gain adaptive change the momentum factor and learning rate. Learning patterns, simulation results indicate that the proposed algorithm can accelerate the convergence behavior and drag the network through deep local minima compared to the conventional BP algorithm. This work focuses upon the training parameters to attain optimal solution for neural network.