Spectral technique for hidden layer neural network training

Terry Windeatt; Robert Tebbs

doi:10.1016/S0167-8655(97)00079-2

Back

Spectral technique for hidden layer neural network training

Journal article

Peer reviewed

Spectral technique for hidden layer neural network training

Terry Windeatt and Robert Tebbs

Pattern recognition letters, Vol.18(8), pp.723-731

01/08/1997

DOI: https://doi.org/10.1016/S0167-8655(97)00079-2

Abstract

k-Monotonic check

Multi-layer perceptrons

Rademacher-Walsh spectrum

Sequential learning systems

Spectrum synthesis

We propose a new constructive algorithm for learning binary-to-binary mappings. Weight constraints derived from a spectral summation are used to check separability during the partitioning phase, and to limit hyperplane movement during training.

Metrics

9 Record Views

8 Times Cited - Web of Science

Details

Title: Spectral technique for hidden layer neural network training
Creators: Terry Windeatt - University of Surrey
Robert Tebbs - University of Surrey
Publication Details: Pattern recognition letters, Vol.18(8), pp.723-731
Publisher: Elsevier B.V
Date published: 01/08/1997
Identifiers: 99782971902346
Academic Unit: School of Computer Science and Electronic Engineering
Language: English
Resource Type: Journal article

Spectral technique for hidden layer neural network training

Abstract

Metrics

Details

Usage Policy