Estimations of error bounds for RBF networks

The training and optimization of neural networks to perform function approximation tasks is well documented in the literature. The usefulness of neural networks will be enhanced if a further capacity is added to them: the ability to estimate the accuracy of the results which they generate. Not only will this provide users of neural networks with a confidence index, it will also enable the estimates from the neural networks to be included as part of an overall estimation scheme in which several estimates are combined in a Bayesian manner to guarantee the optimality (in terms of minimum variance) of the result. For example, it would enable the results from a neural network estimator to be included in a Kalman filter cycle with full mathematical rigour. In this paper the suitability of a perturbation model to perform such a task will be examined.