Boosting the performances of the recurrent neural network by the fuzzy min-max

Ryad Zemouri; Florin Gheorghe Filip; Eugenia Minca; Daniel Racoceanu; Noureddine Zerhouni

Boosting the performances of the recurrent neural network by the fuzzy min-max

Ryad Zemouri
, Florin Gheorghe Filip
, Eugenia Minca
, Daniel Racoceanu
, Noureddine Zerhouni

HESAM Université
Romanian Academy
Valahia University
Agency for Science, Technology and Research, Singapore
CNRS Umr 6174

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

The k-means training algorithm used for the RBF (Radial Basis Function) neural network can have some weakness like empty clusters, the choice of the cluster number and the random choice of the centers of theses clusters. In this paper, we use the Fuzzy Min Max technique to boost the performances of the training algorithm. This technique is used to determine the number of the k centers and to initialize correctly these k centers. The k-means algorithm always converges to the same result for all the tests.

Original language	English
Pages (from-to)	69-90
Number of pages	22
Journal	Romanian Journal of Information Science and Technology
Volume	12
Issue number	1
State	Published - 2009
Externally published	Yes

Keywords

Fuzzy Min-Max
K-means algorithm
RBF network
Recurrent network
Time series prediction

Cite this

@article{5bbf0d70b50f4ef5ab00303ac0d7e06e,

title = "Boosting the performances of the recurrent neural network by the fuzzy min-max",

abstract = "The k-means training algorithm used for the RBF (Radial Basis Function) neural network can have some weakness like empty clusters, the choice of the cluster number and the random choice of the centers of theses clusters. In this paper, we use the Fuzzy Min Max technique to boost the performances of the training algorithm. This technique is used to determine the number of the k centers and to initialize correctly these k centers. The k-means algorithm always converges to the same result for all the tests.",

keywords = "Fuzzy Min-Max, K-means algorithm, RBF network, Recurrent network, Time series prediction",

author = "Ryad Zemouri and Filip, \{Florin Gheorghe\} and Eugenia Minca and Daniel Racoceanu and Noureddine Zerhouni",

year = "2009",

language = "English",

volume = "12",

pages = "69--90",

journal = "Romanian Journal of Information Science and Technology",

issn = "1453-8245",

number = "1",

}

TY - JOUR

T1 - Boosting the performances of the recurrent neural network by the fuzzy min-max

AU - Zemouri, Ryad

AU - Filip, Florin Gheorghe

AU - Minca, Eugenia

AU - Racoceanu, Daniel

AU - Zerhouni, Noureddine

PY - 2009

Y1 - 2009

N2 - The k-means training algorithm used for the RBF (Radial Basis Function) neural network can have some weakness like empty clusters, the choice of the cluster number and the random choice of the centers of theses clusters. In this paper, we use the Fuzzy Min Max technique to boost the performances of the training algorithm. This technique is used to determine the number of the k centers and to initialize correctly these k centers. The k-means algorithm always converges to the same result for all the tests.

AB - The k-means training algorithm used for the RBF (Radial Basis Function) neural network can have some weakness like empty clusters, the choice of the cluster number and the random choice of the centers of theses clusters. In this paper, we use the Fuzzy Min Max technique to boost the performances of the training algorithm. This technique is used to determine the number of the k centers and to initialize correctly these k centers. The k-means algorithm always converges to the same result for all the tests.

KW - Fuzzy Min-Max

KW - K-means algorithm

KW - RBF network

KW - Recurrent network

KW - Time series prediction

UR - https://www.scopus.com/pages/publications/84878638008

M3 - Article

AN - SCOPUS:84878638008

SN - 1453-8245

VL - 12

SP - 69

EP - 90

JO - Romanian Journal of Information Science and Technology

JF - Romanian Journal of Information Science and Technology

IS - 1

ER -

Boosting the performances of the recurrent neural network by the fuzzy min-max

Abstract

Keywords

Other files and links

Fingerprint

Cite this