Simple heuristic approach for training of type-2 NEO-fuzzy neural network

Yancho Todorov; Margarita Terziyska

doi:10.1109/PC.2015.7169976

Simple heuristic approach for training of type-2 NEO-fuzzy neural network

Yancho Todorov
, Margarita Terziyska

Not published at Aalto University

Bulgarian Academy of Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

2 Citations (Scopus)

Abstract

This paper describes the development of Interval Type-2 NEO-Fuzzy Neural Network for modeling of complex dynamics. The proposed network represents a parallel set of multiple zero order Sugeno type approximations, related only to their own input argument. As learning procedure a simple heuristic backpropagation approach, where the sign of the gradient is taken into account, is adopted. To improve the robustness of the network and the possibilities for handling uncertainties, Interval Type-2 Gaussian fuzzy sets are introduced into the network topology. The potentials of the proposed approach in modeling of Mackey-Glass and Rossler Chaotic time series are studied. A comparison is made with the classical Gradient Descent learning approach.

Original language	English
Title of host publication	Proceedings of the 2015 20th International Conference on Process Control, PC 2015
Publisher	IEEE
Pages	278-283
Number of pages	6
Volume	2015-July
ISBN (Electronic)	978-1-4673-6627-4
DOIs	https://doi.org/10.1109/PC.2015.7169976
Publication status	Published - 28 Jul 2015
MoE publication type	A4 Conference publication
Event	International Conference on Process Control - Strbske Pleso, Slovakia Duration: 9 Jun 2015 → 12 Jun 2015 Conference number: 20

Conference

Conference	International Conference on Process Control
Abbreviated title	PC
Country/Territory	Slovakia
City	Strbske Pleso
Period	09/06/2015 → 12/06/2015

Keywords

Chaotic time-series prediction
Dynamic modeling
Fuzzy systems
Neo-fuzzy neuron
Neural networks

Access to Document

10.1109/PC.2015.7169976

Cite this

@inproceedings{ad07a6382ef342f9b238ea13db61bc90,

title = "Simple heuristic approach for training of type-2 NEO-fuzzy neural network",

abstract = "This paper describes the development of Interval Type-2 NEO-Fuzzy Neural Network for modeling of complex dynamics. The proposed network represents a parallel set of multiple zero order Sugeno type approximations, related only to their own input argument. As learning procedure a simple heuristic backpropagation approach, where the sign of the gradient is taken into account, is adopted. To improve the robustness of the network and the possibilities for handling uncertainties, Interval Type-2 Gaussian fuzzy sets are introduced into the network topology. The potentials of the proposed approach in modeling of Mackey-Glass and Rossler Chaotic time series are studied. A comparison is made with the classical Gradient Descent learning approach.",

keywords = "Chaotic time-series prediction, Dynamic modeling, Fuzzy systems, Neo-fuzzy neuron, Neural networks",

author = "Yancho Todorov and Margarita Terziyska",

year = "2015",

month = jul,

day = "28",

doi = "10.1109/PC.2015.7169976",

language = "English",

volume = "2015-July",

pages = "278--283",

booktitle = "Proceedings of the 2015 20th International Conference on Process Control, PC 2015",

publisher = "IEEE",

address = "United States",

note = "International Conference on Process Control, PC ; Conference date: 09-06-2015 Through 12-06-2015",

}

Simple heuristic approach for training of type-2 NEO-fuzzy neural network. / Todorov, Yancho; Terziyska, Margarita.
Proceedings of the 2015 20th International Conference on Process Control, PC 2015. Vol. 2015-July IEEE, 2015. p. 278-283 7169976.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Simple heuristic approach for training of type-2 NEO-fuzzy neural network

AU - Todorov, Yancho

AU - Terziyska, Margarita

N1 - Conference code: 20

PY - 2015/7/28

Y1 - 2015/7/28

N2 - This paper describes the development of Interval Type-2 NEO-Fuzzy Neural Network for modeling of complex dynamics. The proposed network represents a parallel set of multiple zero order Sugeno type approximations, related only to their own input argument. As learning procedure a simple heuristic backpropagation approach, where the sign of the gradient is taken into account, is adopted. To improve the robustness of the network and the possibilities for handling uncertainties, Interval Type-2 Gaussian fuzzy sets are introduced into the network topology. The potentials of the proposed approach in modeling of Mackey-Glass and Rossler Chaotic time series are studied. A comparison is made with the classical Gradient Descent learning approach.

AB - This paper describes the development of Interval Type-2 NEO-Fuzzy Neural Network for modeling of complex dynamics. The proposed network represents a parallel set of multiple zero order Sugeno type approximations, related only to their own input argument. As learning procedure a simple heuristic backpropagation approach, where the sign of the gradient is taken into account, is adopted. To improve the robustness of the network and the possibilities for handling uncertainties, Interval Type-2 Gaussian fuzzy sets are introduced into the network topology. The potentials of the proposed approach in modeling of Mackey-Glass and Rossler Chaotic time series are studied. A comparison is made with the classical Gradient Descent learning approach.

KW - Chaotic time-series prediction

KW - Dynamic modeling

KW - Fuzzy systems

KW - Neo-fuzzy neuron

KW - Neural networks

UR - https://www.scopus.com/pages/publications/84990985649

U2 - 10.1109/PC.2015.7169976

DO - 10.1109/PC.2015.7169976

M3 - Conference article in proceedings

AN - SCOPUS:84990985649

VL - 2015-July

SP - 278

EP - 283

BT - Proceedings of the 2015 20th International Conference on Process Control, PC 2015

PB - IEEE

T2 - International Conference on Process Control

Y2 - 9 June 2015 through 12 June 2015

ER -

Simple heuristic approach for training of type-2 NEO-fuzzy neural network

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this