Authors : Eman Samir Bhaya and Walaa Hussein Ahmed
Abstract: There are many studies on the approximation by neural networks. In general, we cannot approximate a function f in Lp spaces using radial basis function neural networks. We introduce theorems on the degree of best approximation using neural networks for functions defined overall real line. In addition, we define a version of convolution and use it to estimate the degree of approximation using neural networks. Here, we make the approximation using radial basis function neural networks possible by writing some constrains on the target function f. We use the construction in our methods of proofs. Which we can consider them as a function approximation.
Eman Samir Bhaya and Walaa Hussein Ahmed, 2018. Lp, P<1 Approximation Using Radial Basis Function Neural Networks on p Ordered Vector Space. Journal of Engineering and Applied Sciences, 13: 4771-4773.