Some comparisons of radial-basis-function and kernel networks

Capabilities and limitations of two popular computational models, radial-basis-function (RBF) and kernel networks, will be discussed. Higher flexibility in choice of free parameters in RBFs will be compared with benefits of geometrical properties of kernel models. Capabilities of function approximation and generalization will be analysed. General results on scaled kernels will be illustrated by the paradigmatic examples of Gaussian kernel and radial networks.