Cargando…

Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition

In this paper, the method of the graphical interpretation of the single-layer network weights is introduced. It is shown that the network parameters can be converted to the image and their particular elements are the pixels. For this purpose, weight-to-pixel conversion formula is used. Moreover, new...

Descripción completa

Detalles Bibliográficos
Autores principales: Kusy, Maciej, Szczepanski, Damian
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Springer-Verlag 2011
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3442175/
https://www.ncbi.nlm.nih.gov/pubmed/22997482
http://dx.doi.org/10.1007/s00521-011-0754-8
_version_ 1782243439661809664
author Kusy, Maciej
Szczepanski, Damian
author_facet Kusy, Maciej
Szczepanski, Damian
author_sort Kusy, Maciej
collection PubMed
description In this paper, the method of the graphical interpretation of the single-layer network weights is introduced. It is shown that the network parameters can be converted to the image and their particular elements are the pixels. For this purpose, weight-to-pixel conversion formula is used. Moreover, new weights’ modification method is proposed. The weight coefficients are computed on the basis of pixel values for which image filtration algorithms are implemented. The approach is applied to the weights of three types of the models: single-layer network, two-layer backpropagation network and the hybrid network. The performance of the models is then compared on two independent data sets. By means of the experiments, it is presented that the adjustment of the weights to new values decreases test error value compared to the error obtained for initial set of weights.
format Online
Article
Text
id pubmed-3442175
institution National Center for Biotechnology Information
language English
publishDate 2011
publisher Springer-Verlag
record_format MEDLINE/PubMed
spelling pubmed-34421752012-09-18 Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition Kusy, Maciej Szczepanski, Damian Neural Comput Appl Original Article In this paper, the method of the graphical interpretation of the single-layer network weights is introduced. It is shown that the network parameters can be converted to the image and their particular elements are the pixels. For this purpose, weight-to-pixel conversion formula is used. Moreover, new weights’ modification method is proposed. The weight coefficients are computed on the basis of pixel values for which image filtration algorithms are implemented. The approach is applied to the weights of three types of the models: single-layer network, two-layer backpropagation network and the hybrid network. The performance of the models is then compared on two independent data sets. By means of the experiments, it is presented that the adjustment of the weights to new values decreases test error value compared to the error obtained for initial set of weights. Springer-Verlag 2011-11-11 2012 /pmc/articles/PMC3442175/ /pubmed/22997482 http://dx.doi.org/10.1007/s00521-011-0754-8 Text en © The Author(s) 2011 https://creativecommons.org/licenses/by-nc/4.0/ This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
spellingShingle Original Article
Kusy, Maciej
Szczepanski, Damian
Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title_full Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title_fullStr Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title_full_unstemmed Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title_short Influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
title_sort influence of graphical weights’ interpretation and filtration algorithms on generalization ability of neural networks applied to digit recognition
topic Original Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3442175/
https://www.ncbi.nlm.nih.gov/pubmed/22997482
http://dx.doi.org/10.1007/s00521-011-0754-8
work_keys_str_mv AT kusymaciej influenceofgraphicalweightsinterpretationandfiltrationalgorithmsongeneralizationabilityofneuralnetworksappliedtodigitrecognition
AT szczepanskidamian influenceofgraphicalweightsinterpretationandfiltrationalgorithmsongeneralizationabilityofneuralnetworksappliedtodigitrecognition