A New Improved BP Neural Network Algorithm

Neural network is widely used in pattern recognition, image processing and system control. BP neural network has its inherent deficiencies. Its convergence rate is slow. It is easy to fall into the local minimum and the structure of the neural network is hard to determine. The structure of hidden layer is determined through the experience, but it can not make accurate judgments with complex network structure. In order to improve the function of the BP neural network, an improved algorithm of BP neural network based on the standard sigmoid function is put forward. fuzzy theory is added to the algorithm to determine the structure of hidden layer and dynamically adjusted additional momentum factor is also added. Compare with conventional algorithms it has a greater ability to enhance the study, reduce the hidden layers´ nodes effectively, and it also has a higher network convergence speed and precision.