Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Regularised feed forward neural networks for streamed data classification problems
Dissertation (MSc)--University of Pretoria, 2020.
Saved in:
| Other Authors: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
University of Pretoria
2020
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613544324792320 |
|---|---|
| access_status_str | Open Access |
| author2 | Bosman, Anna S. |
| author_browse | Bosman, Anna S. |
| author_facet | Bosman, Anna S. |
| collection | Thesis |
| dc_rights_str_mv | © 2019 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, 2020. |
| format | Thesis |
| id | oai:repository.up.ac.za:2263/75804 |
| institution | University of Pretoria (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:37:50.174Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from UPSpace — University of Pretoria Institutional Repository |
| publishDate | 2020 |
| publishDateRange | 2020 |
| publishDateSort | 2020 |
| 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:2263/75804 Regularised feed forward neural networks for streamed data classification problems Bosman, Anna S. mox.1990@gmail.com Engelbrecht, Andries P. Ellis, Mathys Computational Intelligence Data Streams Feed Forward Neural Networks Quantum Particle Swarm Optimisation Regularisation Feed Forward Neural Networks Classification Problems Concept drift UCTD Engineering, built environment and information technology theses SDG-08 Engineering, built environment and information technology theses SDG-09 Dissertation (MSc)--University of Pretoria, 2020. Streamed data classification problems (SDCPs) require classifiers with the ability to learn and to adjust to the underlying relationships in data streams, in real-time. This requirement poses a challenge to classifiers, because the learning task is no longer just to find the optimal decision boundaries, but also to track changes in the decision boundaries as new training data is received. The challenge is due to concept drift, i.e. the changing of decision boundaries over time. Changes include disappearing, appearing, or shifting decision boundaries. This thesis proposes an online learning approach for feed forward neural networks (FFNNs) that meets the requirements of SDCPs. The approach uses regularisation to optimise the architecture via the weights, and quantum particle swarm optimisation (QPSO) to dynamically adjust the weights. The learning approach is applied to a FFNN, which uses rectified linear activation functions, to form a novel SDCP classifier. The classifier is empirically investigated on several SDCPs. Both weight decay (WD) and weight elimination (WE) are investigated as regularisers. Empirical results show that using QPSO with no regularisation, causes the classifier to completely saturate. However, using QPSO with regularisation enables the classifier to dynamically adapt both its implicit architecture and weights as decision boundaries change. Furthermore, the results favour WE over WD as a regulariser for QPSO. National Research Foundation (NRF) bs2026 Computer Science MSc Unrestricted SDG-08: Decent work and economic growth SDG-09: Industry, innovation and infrastructure 2020-08-19T08:06:14Z 2020-08-19T08:06:14Z 2020-09 2020 Dissertation Ellis, M 2020, Regularised feed forward neural networks for streamed data classification problems, MSc Dissertation, University of Pretoria, Pretoria, viewed yymmdd <http://hdl.handle.net/2263/75804> S2020 http://hdl.handle.net/2263/75804 en © 2019 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 Data Streams Feed Forward Neural Networks Quantum Particle Swarm Optimisation Regularisation Feed Forward Neural Networks Classification Problems Concept drift UCTD Engineering, built environment and information technology theses SDG-08 Engineering, built environment and information technology theses SDG-09 Regularised feed forward neural networks for streamed data classification problems |
| title | Regularised feed forward neural networks for streamed data classification problems |
| title_full | Regularised feed forward neural networks for streamed data classification problems |
| title_fullStr | Regularised feed forward neural networks for streamed data classification problems |
| title_full_unstemmed | Regularised feed forward neural networks for streamed data classification problems |
| title_short | Regularised feed forward neural networks for streamed data classification problems |
| title_sort | regularised feed forward neural networks for streamed data classification problems |
| topic | Computational Intelligence Data Streams Feed Forward Neural Networks Quantum Particle Swarm Optimisation Regularisation Feed Forward Neural Networks Classification Problems Concept drift UCTD Engineering, built environment and information technology theses SDG-08 Engineering, built environment and information technology theses SDG-09 |
| url | http://hdl.handle.net/2263/75804 |