Toggle Main Menu Toggle Search

Open Access padlockePrints

A training method for enhancing neural network model generalisation

Lookup NU author(s): Dr Jie ZhangORCiD

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Abstract

A training method for enhancing neural network model generalisation is proposed in this paper. In this method, a neural network is trained and tested alternatively on a training data set and a testing data set. Unlike in conventional neural network training where the training and testing data sets are fixed, the training and testing data sets swap roles continuously during network training. Training is terminated when the network prediction errors on both data sets cannot be further reduced. Application examples demonstrate that this neural network training strategy can significantly improve neural network model prediction accuracy, especially long range prediction accuracy.


Publication metadata

Author(s): Zhang J

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: International Joint Conference on Neural Networks

Year of Conference: 2002

Pages: 800-805

Publisher: IEEE

URL: http://dx.doi.org/10.1109/IJCNN.2002.1005576

DOI: 10.1109/IJCNN.2002.1005576

Library holdings: Search Newcastle University Library for this item

ISBN: 0780372786


Share