Reduced rule-base fuzzy-neural networks

Margarita Terziyska, Yancho Todorov

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

In this paper two different fuzzy-neural systems with reduced fuzzy rules bases, namely Distributed Adaptive Neuro Fuzzy Architecture (DANFA) and Semi Fuzzy Neural Network (SFNN), are presented. Both structures are realized with Takagi-Sugeno fuzzy inference mechanism and they posses reduced number of parameters for update during the learning procedure. Thus, the computational time for algorithm execution is additionally reduced, which make the modeling structures a promising solution for real time applications. As a learning approach for the designed structures a simplified two-step gradient descent approach is implemented. To demonstrate the potentials of both models, simulation experiments with two benchmark chaotic time systems—Mackey-Glass and Rossler are studied. The obtained results show accurate models performance with minimal prediction error.

Original languageEnglish
Title of host publicationAdvanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM
EditorsIvan Georgiev, Michail Todorov, Krassimir Georgiev
Pages199-214
Number of pages16
DOIs
Publication statusPublished - 1 Jan 2017
MoE publication typeA4 Article in a conference publication
Event10th Annual Meeting of the Bulgarian Section of SIAM, BGSIAM 2015 - Sofia, Bulgaria
Duration: 21 Dec 201522 Dec 2015

Publication series

NameStudies in Computational Intelligence
Volume681
ISSN (Print)1860-949X

Conference

Conference10th Annual Meeting of the Bulgarian Section of SIAM, BGSIAM 2015
CountryBulgaria
City Sofia
Period21/12/1522/12/15

Fingerprint

Fuzzy neural networks
Fuzzy inference
Fuzzy rules
Glass
Experiments

Cite this

Terziyska, M., & Todorov, Y. (2017). Reduced rule-base fuzzy-neural networks. In I. Georgiev, M. Todorov, & K. Georgiev (Eds.), Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM (pp. 199-214). Studies in Computational Intelligence, Vol.. 681 https://doi.org/10.1007/978-3-319-49544-6_17
Terziyska, Margarita ; Todorov, Yancho. / Reduced rule-base fuzzy-neural networks. Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM. editor / Ivan Georgiev ; Michail Todorov ; Krassimir Georgiev. 2017. pp. 199-214 (Studies in Computational Intelligence, Vol. 681).
@inproceedings{2ed9cfdee8ad450989fb8082fcd721da,
title = "Reduced rule-base fuzzy-neural networks",
abstract = "In this paper two different fuzzy-neural systems with reduced fuzzy rules bases, namely Distributed Adaptive Neuro Fuzzy Architecture (DANFA) and Semi Fuzzy Neural Network (SFNN), are presented. Both structures are realized with Takagi-Sugeno fuzzy inference mechanism and they posses reduced number of parameters for update during the learning procedure. Thus, the computational time for algorithm execution is additionally reduced, which make the modeling structures a promising solution for real time applications. As a learning approach for the designed structures a simplified two-step gradient descent approach is implemented. To demonstrate the potentials of both models, simulation experiments with two benchmark chaotic time systems—Mackey-Glass and Rossler are studied. The obtained results show accurate models performance with minimal prediction error.",
author = "Margarita Terziyska and Yancho Todorov",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-49544-6_17",
language = "English",
isbn = "978-3-31949543-9",
series = "Studies in Computational Intelligence",
pages = "199--214",
editor = "Ivan Georgiev and Michail Todorov and Krassimir Georgiev",
booktitle = "Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM",

}

Terziyska, M & Todorov, Y 2017, Reduced rule-base fuzzy-neural networks. in I Georgiev, M Todorov & K Georgiev (eds), Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM. Studies in Computational Intelligence, vol. 681, pp. 199-214, 10th Annual Meeting of the Bulgarian Section of SIAM, BGSIAM 2015, Sofia, Bulgaria, 21/12/15. https://doi.org/10.1007/978-3-319-49544-6_17

Reduced rule-base fuzzy-neural networks. / Terziyska, Margarita; Todorov, Yancho.

Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM. ed. / Ivan Georgiev; Michail Todorov; Krassimir Georgiev. 2017. p. 199-214 (Studies in Computational Intelligence, Vol. 681).

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Reduced rule-base fuzzy-neural networks

AU - Terziyska, Margarita

AU - Todorov, Yancho

PY - 2017/1/1

Y1 - 2017/1/1

N2 - In this paper two different fuzzy-neural systems with reduced fuzzy rules bases, namely Distributed Adaptive Neuro Fuzzy Architecture (DANFA) and Semi Fuzzy Neural Network (SFNN), are presented. Both structures are realized with Takagi-Sugeno fuzzy inference mechanism and they posses reduced number of parameters for update during the learning procedure. Thus, the computational time for algorithm execution is additionally reduced, which make the modeling structures a promising solution for real time applications. As a learning approach for the designed structures a simplified two-step gradient descent approach is implemented. To demonstrate the potentials of both models, simulation experiments with two benchmark chaotic time systems—Mackey-Glass and Rossler are studied. The obtained results show accurate models performance with minimal prediction error.

AB - In this paper two different fuzzy-neural systems with reduced fuzzy rules bases, namely Distributed Adaptive Neuro Fuzzy Architecture (DANFA) and Semi Fuzzy Neural Network (SFNN), are presented. Both structures are realized with Takagi-Sugeno fuzzy inference mechanism and they posses reduced number of parameters for update during the learning procedure. Thus, the computational time for algorithm execution is additionally reduced, which make the modeling structures a promising solution for real time applications. As a learning approach for the designed structures a simplified two-step gradient descent approach is implemented. To demonstrate the potentials of both models, simulation experiments with two benchmark chaotic time systems—Mackey-Glass and Rossler are studied. The obtained results show accurate models performance with minimal prediction error.

UR - http://www.scopus.com/inward/record.url?scp=85012247629&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-49544-6_17

DO - 10.1007/978-3-319-49544-6_17

M3 - Conference article in proceedings

SN - 978-3-31949543-9

T3 - Studies in Computational Intelligence

SP - 199

EP - 214

BT - Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM

A2 - Georgiev, Ivan

A2 - Todorov, Michail

A2 - Georgiev, Krassimir

ER -

Terziyska M, Todorov Y. Reduced rule-base fuzzy-neural networks. In Georgiev I, Todorov M, Georgiev K, editors, Advanced Computing in Industrial Mathematics - Revised Selected Papers of the 10th Annual Meeting of the Bulgarian Section of SIAM. 2017. p. 199-214. (Studies in Computational Intelligence, Vol. 681). https://doi.org/10.1007/978-3-319-49544-6_17