Determining functional relationships from trained neural networks

Authors:
A. M. Hammitt;E. B. Bartlett
Affiliations:
Iowa State University Ames, IA, 50011-2241, U.S.A.;Iowa State University Ames, IA, 50011-2241, U.S.A.
Venue:
Mathematical and Computer Modelling: An International Journal
Year:
1995

Citing 15
Cited 0

Numerical recipes: the art of scientific computing

Numerical recipes: the art of scientific computing
Monte Carlo optimization, simulation, and sensitivity of queueing networks

Monte Carlo optimization, simulation, and sensitivity of queueing networks
The ART of Adaptive Pattern Recognition by a Self-Organizing Neural Network

Computer
Multilayer feedforward networks are universal approximators

Neural Networks
A new approach for finding the global minimum of error function* of neural networks

Neural Networks
On the approximate realization of continuous mappings by neural networks

Neural Networks
The computation of generalized cross-validation functions through householder tridiagonalization with applications to the fitting of interaction spline models

SIAM Journal on Matrix Analysis and Applications
Neural network design and the complexity of learning

Neural network design and the complexity of learning
What size net gives valid generalization?

Neural Computation
Connectionist nonparametric regression: multilayer feedforward networks can learn arbitrary mappings

Neural Networks
Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks

Neural Networks
The II &Pgr; method for estimating multivariate functions from noisy data

Technometrics
Interpreting neural-network connection weights

AI Expert
Neural nets tell why

Dr. Dobb's Journal
Original Contribution: On the complexity of approximating mappings using feedforward networks

Neural Networks

Quantified Score

Hi-index	0.98

Visualization

Abstract

Research has shown that artificial neural networks (ANNs) can be trained to perform function mapping tasks. In this work, ANN mappings are approximated around an operating point by third-order polynomials (3OPs). A previously unknown function can be modeled by an ANN, and a 3OP can be derived from the ANN model. In addition, the sensitivity of an output to changes in the inputs can be easily determined from these polynomials in large neighborhoods around the operating points. Examples are shown that illustrate the use of the ANN-3OP mapping approach.