Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Using particle swarm optimisation to train feedforward neural networks in dynamic environments
Dissertation (MSc)--University of Pretoria, 2011.
Saved in:
| Other Authors: | |
|---|---|
| Format: | Thesis |
| Published: |
University of Pretoria
2013
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613516463079424 |
|---|---|
| access_status_str | Open Access |
| author2 | Engelbrecht, Andries P. |
| author_browse | Engelbrecht, Andries P. |
| author_facet | Engelbrecht, Andries P. |
| collection | Thesis |
| dc_rights_str_mv | © 2012, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. |
| description | Dissertation (MSc)--University of Pretoria, 2011. |
| format | Thesis |
| id | oai:repository.up.ac.za:2263/28618 |
| institution | University of Pretoria (South Africa) |
| last_indexed | 2026-06-10T12:37:23.532Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from UPSpace — University of Pretoria Institutional Repository |
| publishDate | 2013 |
| publishDateRange | 2013 |
| publishDateSort | 2013 |
| publisher | University of Pretoria |
| publisherStr | University of Pretoria |
| record_format | dspace |
| source_str | UPSpace — University of Pretoria Institutional Repository |
| spelling | oai:repository.up.ac.za:2263/28618 Using particle swarm optimisation to train feedforward neural networks in dynamic environments Engelbrecht, Andries P. myearwen@gmail.com Rakitianskaia, A.S. (Anastassia Sergeevna) Computational intelligence Particle swarm optimization (PSO) Concept drift Neural networks UCTD Dissertation (MSc)--University of Pretoria, 2011. The feedforward neural network (NN) is a mathematical model capable of representing any non-linear relationship between input and output data. It has been succesfully applied to a wide variety of classification and function approximation problems. Various neural network training algorithms were developed, including the particle swarm optimiser (PSO), which was shown to outperform the standard back propagation training algorithm on a selection of problems. However, it was usually assumed that the environment in which a NN operates is static. Such an assumption is often not valid for real life problems, and the training algorithms have to be adapted accordingly. Various dynamic versions of the PSO have already been developed. This work investigates the applicability of dynamic PSO algorithms to NN training in dynamic environments, and compares the performance of dynamic PSO algorithms to the performance of back propagation. Three popular dynamic PSO variants are considered. The extent of adaptive properties of back propagation and dynamic PSO under different kinds of dynamic environments is determined. Dynamic PSO is shown to be a viable alternative to back propagation, especially under the environments exhibiting infrequent gradual changes. Copyright 2011, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. Please cite as follows: Rakitianskaia, A 2011, Using particle swarm optimisation to train feedforward neural networks in dynamic environments, MSc dissertation, University of Pretoria, Pretoria, viewed yymmdd < http://upetd.up.ac.za/thesis/available/etd-02132012-233212 / > C12/4/406/gm Computer Science Unrestricted 2013-09-07T13:50:19Z 2012-05-02 2013-09-07T13:50:19Z 2012-04-19 2011 2012-02-13 Dissertation Rakitianskaia, A 2011, Using particle swarm optimisation to train feedforward neural networks in dynamic environments, MSc dissertation, University of Pretoria, Pretoria, viewed yymmdd < http://hdl.handle.net/2263/28618 > http://hdl.handle.net/2263/28618 http://upetd.up.ac.za/thesis/available/etd-02132012-233212/ © 2012, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. application/pdf University of Pretoria |
| spellingShingle | Computational intelligence Particle swarm optimization (PSO) Concept drift Neural networks UCTD Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title | Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title_full | Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title_fullStr | Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title_full_unstemmed | Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title_short | Using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| title_sort | using particle swarm optimisation to train feedforward neural networks in dynamic environments |
| topic | Computational intelligence Particle swarm optimization (PSO) Concept drift Neural networks UCTD |
| url | http://hdl.handle.net/2263/28618 http://upetd.up.ac.za/thesis/available/etd-02132012-233212/ |