Fuzzy neural networks stability in terms of the number of hidden layers

R. Lovassy, L. Kóczy, L. Gál, I. Rudas

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

This paper introduces an approach for studying the stability, and generalization capability of one and two hidden layer Fuzzy Flip-Flop based Neural Networks (FNNs) with various fuzzy operators. By employing fuzzy flip-flop neurons as sigmoid function generators, novel function approximators are established that also avoid overfitting in the case of test data containing noisy items in the form of outliers. It is shown, by comparing with existing standard tansig function based approaches that reducing the network complexity networks with comparable stability are obtained. Finally, examples are given to illustrate the effect of the hidden layer number of neural networks.

Original languageEnglish
Title of host publication12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings
Pages323-328
Number of pages6
DOIs
Publication statusPublished - 2011
Event12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Budapest, Hungary
Duration: Nov 21 2011Nov 22 2011

Other

Other12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011
CountryHungary
CityBudapest
Period11/21/1111/22/11

Fingerprint

Flip flop circuits
Fuzzy neural networks
Function generators
Neural networks
Neurons

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computational Theory and Mathematics
  • Information Systems

Cite this

Lovassy, R., Kóczy, L., Gál, L., & Rudas, I. (2011). Fuzzy neural networks stability in terms of the number of hidden layers. In 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings (pp. 323-328). [6108523] https://doi.org/10.1109/CINTI.2011.6108523

Fuzzy neural networks stability in terms of the number of hidden layers. / Lovassy, R.; Kóczy, L.; Gál, L.; Rudas, I.

12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings. 2011. p. 323-328 6108523.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lovassy, R, Kóczy, L, Gál, L & Rudas, I 2011, Fuzzy neural networks stability in terms of the number of hidden layers. in 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings., 6108523, pp. 323-328, 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011, Budapest, Hungary, 11/21/11. https://doi.org/10.1109/CINTI.2011.6108523
Lovassy R, Kóczy L, Gál L, Rudas I. Fuzzy neural networks stability in terms of the number of hidden layers. In 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings. 2011. p. 323-328. 6108523 https://doi.org/10.1109/CINTI.2011.6108523
Lovassy, R. ; Kóczy, L. ; Gál, L. ; Rudas, I. / Fuzzy neural networks stability in terms of the number of hidden layers. 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings. 2011. pp. 323-328
@inproceedings{59e3a064324546e39b6821f2fe758430,
title = "Fuzzy neural networks stability in terms of the number of hidden layers",
abstract = "This paper introduces an approach for studying the stability, and generalization capability of one and two hidden layer Fuzzy Flip-Flop based Neural Networks (FNNs) with various fuzzy operators. By employing fuzzy flip-flop neurons as sigmoid function generators, novel function approximators are established that also avoid overfitting in the case of test data containing noisy items in the form of outliers. It is shown, by comparing with existing standard tansig function based approaches that reducing the network complexity networks with comparable stability are obtained. Finally, examples are given to illustrate the effect of the hidden layer number of neural networks.",
author = "R. Lovassy and L. K{\'o}czy and L. G{\'a}l and I. Rudas",
year = "2011",
doi = "10.1109/CINTI.2011.6108523",
language = "English",
isbn = "9781457700453",
pages = "323--328",
booktitle = "12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings",

}

TY - GEN

T1 - Fuzzy neural networks stability in terms of the number of hidden layers

AU - Lovassy, R.

AU - Kóczy, L.

AU - Gál, L.

AU - Rudas, I.

PY - 2011

Y1 - 2011

N2 - This paper introduces an approach for studying the stability, and generalization capability of one and two hidden layer Fuzzy Flip-Flop based Neural Networks (FNNs) with various fuzzy operators. By employing fuzzy flip-flop neurons as sigmoid function generators, novel function approximators are established that also avoid overfitting in the case of test data containing noisy items in the form of outliers. It is shown, by comparing with existing standard tansig function based approaches that reducing the network complexity networks with comparable stability are obtained. Finally, examples are given to illustrate the effect of the hidden layer number of neural networks.

AB - This paper introduces an approach for studying the stability, and generalization capability of one and two hidden layer Fuzzy Flip-Flop based Neural Networks (FNNs) with various fuzzy operators. By employing fuzzy flip-flop neurons as sigmoid function generators, novel function approximators are established that also avoid overfitting in the case of test data containing noisy items in the form of outliers. It is shown, by comparing with existing standard tansig function based approaches that reducing the network complexity networks with comparable stability are obtained. Finally, examples are given to illustrate the effect of the hidden layer number of neural networks.

UR - http://www.scopus.com/inward/record.url?scp=84855941334&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84855941334&partnerID=8YFLogxK

U2 - 10.1109/CINTI.2011.6108523

DO - 10.1109/CINTI.2011.6108523

M3 - Conference contribution

AN - SCOPUS:84855941334

SN - 9781457700453

SP - 323

EP - 328

BT - 12th IEEE International Symposium on Computational Intelligence and Informatics, CINTI 2011 - Proceedings

ER -