Approximation of continuous functions defined on a curve by feed-forward neural networks.

It's an important problem in computer aided geometry design to approximate continuous functions defined on a smooth curve, which is widely applied in industry and engineering. In this paper, we approximate continuous functions defined on a smooth single-value curve by means of a new and popular tool- single-hidden layer feed-forward neural networks. It's checked that the approximation is simple, quick and can be easily realized by the Matlab.