Neural networks and nonparametric regression

Abstract

The problem of estimating an unknown function from a finite number of noisy data points is a problem of fundamental importance for many applications in signal processing, machine vision, pattern recognition, and process control. Recently, several new computational techniques for nonparametric regression have been proposed by statisticians and by researchers in artificial neural networks. The author presents a critical survey and a common taxonomy of statistical and neural network methods for regression. Global parametric methods, piecewise parametric or locally parametric methods, and adaptive computation methods are considered. >