In this paper, the interpolation mechanism of functional networks is discussed. A kind of fourlayer
(with 1 input and 1 output unit) and a five-layer (with double input and single output unit)
functional network are designed. Meanwhile, a learning algorithm based on Minimizing a least
squares error function with a unique minimum has been proposed for the purpose of
approximating function. Experimental results indicate that the interpolation method is
completely correct.ÿ ['˜
['