Improving neural networks training using experiment design approach

This project involves the use of Neural Networks (NN) for function approximation. Conventionally, the parameters of a neural network are tuned by minimizing an objective function based on a pre-determined set of training data. This training paradigm is passive in the sense that the neural network on...

Full description

Saved in:
Bibliographic Details
Main Author: Chong, Wei Kean
Format: Thesis
Language:English
Published: 2005
Subjects:
Online Access:http://eprints.utm.my/id/eprint/3580/2/ChongWeiKeanMED2005.PDF
http://eprints.utm.my/id/eprint/3580/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utm.3580
record_format eprints
spelling my.utm.35802020-03-09T07:02:49Z http://eprints.utm.my/id/eprint/3580/ Improving neural networks training using experiment design approach Chong, Wei Kean TK Electrical engineering. Electronics Nuclear engineering This project involves the use of Neural Networks (NN) for function approximation. Conventionally, the parameters of a neural network are tuned by minimizing an objective function based on a pre-determined set of training data. This training paradigm is passive in the sense that the neural network only learns from the training patterns presented to it by the environment or a teacher. It may be more useful if the neural network could 'actively' obtain the training samples itself sequentially by interacting with its environment. There are several methods of selecting training data from input space for neural networks which include D-optimal and Max-min design approaches. Consider a function approximation problem (Neural Network using Radial Basic Function structure) and limit the amount of training data, say (m) from N amount of possible data. Randomly select the m data set for conventional training algorithm. One more data (m+ 1) is entered to train the NN again. This data is selected by two methods: random and Experiment Design Approach (D-optimal and Maxmin Distance). The performances of the two approaches are then compared. It was found that the NN trained using the data obtained using experiment design approaches approximated the unknown function better than the NN that is trained when the data are selected randomly. Maxmin Distance Approach is independent of the NN model while Doptimal point is dependent on the NN model used. 2005-03 Thesis NonPeerReviewed application/pdf en http://eprints.utm.my/id/eprint/3580/2/ChongWeiKeanMED2005.PDF Chong, Wei Kean (2005) Improving neural networks training using experiment design approach. Masters thesis, Universiti Teknologi Malaysia, Faculty of Electrical Engineering.
institution Universiti Teknologi Malaysia
building UTM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknologi Malaysia
content_source UTM Institutional Repository
url_provider http://eprints.utm.my/
language English
topic TK Electrical engineering. Electronics Nuclear engineering
spellingShingle TK Electrical engineering. Electronics Nuclear engineering
Chong, Wei Kean
Improving neural networks training using experiment design approach
description This project involves the use of Neural Networks (NN) for function approximation. Conventionally, the parameters of a neural network are tuned by minimizing an objective function based on a pre-determined set of training data. This training paradigm is passive in the sense that the neural network only learns from the training patterns presented to it by the environment or a teacher. It may be more useful if the neural network could 'actively' obtain the training samples itself sequentially by interacting with its environment. There are several methods of selecting training data from input space for neural networks which include D-optimal and Max-min design approaches. Consider a function approximation problem (Neural Network using Radial Basic Function structure) and limit the amount of training data, say (m) from N amount of possible data. Randomly select the m data set for conventional training algorithm. One more data (m+ 1) is entered to train the NN again. This data is selected by two methods: random and Experiment Design Approach (D-optimal and Maxmin Distance). The performances of the two approaches are then compared. It was found that the NN trained using the data obtained using experiment design approaches approximated the unknown function better than the NN that is trained when the data are selected randomly. Maxmin Distance Approach is independent of the NN model while Doptimal point is dependent on the NN model used.
format Thesis
author Chong, Wei Kean
author_facet Chong, Wei Kean
author_sort Chong, Wei Kean
title Improving neural networks training using experiment design approach
title_short Improving neural networks training using experiment design approach
title_full Improving neural networks training using experiment design approach
title_fullStr Improving neural networks training using experiment design approach
title_full_unstemmed Improving neural networks training using experiment design approach
title_sort improving neural networks training using experiment design approach
publishDate 2005
url http://eprints.utm.my/id/eprint/3580/2/ChongWeiKeanMED2005.PDF
http://eprints.utm.my/id/eprint/3580/
_version_ 1662754213225037824
score 13.209306