New algorithm on neural networks using Padé weight functions

A new algorithm using Padé weight function for training neural networks is proposed in this paper with simple network topology constituted by two layers: input layer and output layer. The process of how to get the Padé weight function networks from the sample interpolation points is given. The new algorithm proposed in this paper can avoid several disadvantages such as local minimum, slow learning speed, and difficulty in obtaining of global optimal point in traditional neural network´s models, Simulation examples show the good performance of the new algorithm with high accuracy and learning speed.