Pointwise approximation for neural networks

It is shown in this paper by a constructive method that for anY f ε C^(m) [a, b], the function and its m order derivatives can be simultaneously approximated by a neural network with one hidden layer in the pointwise sense. This approach naturally yields the design of the hidden layer and the estimate of rate of convergence. The obtained results describe the relationship among the approximation degree of networks, the number of neurons in the hidden layer and the input sample, and reveal that the approximation speed of the constructed networks depends on the smoothness of approximated function.