Projection pursuit learning

Abstract
A learning model based on a nonparametric statistical technique, projection pursuit regression, is studied. Projection pursuit is a nonparametric statistical technique to find interesting low-dimensional projections of high-dimensional data sets. Projection pursuit regression approximates a function of q variables by a sum of nonlinear functions of linear combinations of the q variables, which is related to current neural network models. A training algorithm for projection pursuit learning, called backfitting, is investigated. An example of the application of this model is demonstrated.

This publication has 5 references indexed in Scilit: