Function approximation using neural networks: a simulation study
Date
1992
Authors
Contributor
Advisor
Department
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Narrator
Transcriber
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
University of Hawaii at Manoa
Volume
Number/Issue
Starting Page
Ending Page
Alternative Title
Abstract
This dissertation analyzes the effect of different characteristics of data on the training and estimation accuracy of neural networks. The literature on the universal approximation property of neural networks is reviewed. An examination of the relationship of the neural network approach to traditional statistical methods of approximation brought about proposed enhancements to the neural network training procedure. The study generated data samples characterized by different functional forms, levels of random noise, number and magnitude of outliers, and strength of multicollinearity. These samples were then used to train a neural network. The accuracy of the neural network estimate was tested and compared with the accuracy of the estimates obtained from the true model and those from Specht's GRNN model. Statistics on the length of training and the complexity of the neural network estimate were also collected and analyzed.
Description
Keywords
Citation
Extent
Format
Geographic Location
Time Period
Related To
Theses for the degree of Doctor of Philosophy (University of Hawaii at Manoa). Communication and Information Sciences; no. 2782
Related To (URI)
Table of Contents
Rights
All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.