On a multiobjective training algorithm for RBF networks using particle swarm optimization

G. R L Silva; D. A G Vieira; A. C. Lisboa; Vasile Palade

doi:10.1109/ICTAI.2010.112

On a multiobjective training algorithm for RBF networks using particle swarm optimization

G. R L Silva, D. A G Vieira, A. C. Lisboa, Vasile Palade

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

Abstract

This paper presents a novel algorithm for multiobjective training of Radial Basis Function (RBF) networks based on least-squares and Particle Swarm Optimization methods. The formulation is based on the fundamental concept that supervised learning is a bi-objective optimization problem, in which two conflicting objectives should be minimized. The objectives are related to the empirical training error and the machine complexity. The training is done in three steps: i) a conventional minimization of the training error, ii) multiobjective least-squares optimization for the linear parameters and, iii) particle swarm optimization for the nonlinear parameters. Some results are presented and they show the effectiveness of the proposed approach.

Original language	English
Title of host publication	Proceedings - 22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010
Pages	282-285
Number of pages	4
Volume	2
ISBN (Electronic)	978-0-7695-4263-8
DOIs	https://doi.org/10.1109/ICTAI.2010.112
Publication status	Published - 17 Dec 2010
Externally published	Yes
Event	22nd International Conference on Tools with Artificial Intelligence - Arras, France Duration: 27 Oct 2010 → 29 Oct 2010

Conference

Conference	22nd International Conference on Tools with Artificial Intelligence
Abbreviated title	ICTAI 2010
Country/Territory	France
City	Arras
Period	27/10/10 → 29/10/10

Keywords

Training
Radial basis function networks
Complexity theory
Particle swarm optimization
Artificial neural networks
Optimization
Machine learning
radial basis function networks
learning (artificial intelligence)
least squares approximations
particle swarm optimisation
linear parameters
multiobjective training
RBF networks
particle swarm optimization
least-squares approximation
supervised learning
biobjective optimization problem
machine complexity
multiobjective least-squares optimization
radial basis network
multiobjective least squares

ASJC Scopus subject areas

Software
Artificial Intelligence
Computer Science Applications

Access to Document

10.1109/ICTAI.2010.112

Cite this

On a multiobjective training algorithm for RBF networks using particle swarm optimization. / Silva, G. R L; Vieira, D. A G; Lisboa, A. C. et al.
Proceedings - 22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010. Vol. 2 2010. p. 282-285 5671402.

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

Silva, GRL, Vieira, DAG, Lisboa, AC & Palade, V 2010, On a multiobjective training algorithm for RBF networks using particle swarm optimization. in Proceedings - 22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010. vol. 2, 5671402, pp. 282-285, 22nd International Conference on Tools with Artificial Intelligence, Arras, France, 27/10/10. https://doi.org/10.1109/ICTAI.2010.112

@inproceedings{d79e3b3a6d214d4ba2e60a4ea18aaade,

title = "On a multiobjective training algorithm for RBF networks using particle swarm optimization",

abstract = "This paper presents a novel algorithm for multiobjective training of Radial Basis Function (RBF) networks based on least-squares and Particle Swarm Optimization methods. The formulation is based on the fundamental concept that supervised learning is a bi-objective optimization problem, in which two conflicting objectives should be minimized. The objectives are related to the empirical training error and the machine complexity. The training is done in three steps: i) a conventional minimization of the training error, ii) multiobjective least-squares optimization for the linear parameters and, iii) particle swarm optimization for the nonlinear parameters. Some results are presented and they show the effectiveness of the proposed approach.",

keywords = "Training, Radial basis function networks, Complexity theory, Particle swarm optimization, Artificial neural networks, Optimization, Machine learning, radial basis function networks, learning (artificial intelligence), least squares approximations, particle swarm optimisation, linear parameters, multiobjective training, RBF networks, particle swarm optimization, least-squares approximation, supervised learning, biobjective optimization problem, machine complexity, multiobjective least-squares optimization, radial basis network, multiobjective least squares",

author = "Silva, {G. R L} and Vieira, {D. A G} and Lisboa, {A. C.} and Vasile Palade",

year = "2010",

month = dec,

day = "17",

doi = "10.1109/ICTAI.2010.112",

language = "English",

isbn = "978-1-4244-8817-9",

volume = "2",

pages = "282--285",

booktitle = "Proceedings - 22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010",

note = "22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010 ; Conference date: 27-10-2010 Through 29-10-2010",

}

TY - GEN

T1 - On a multiobjective training algorithm for RBF networks using particle swarm optimization

AU - Silva, G. R L

AU - Vieira, D. A G

AU - Lisboa, A. C.

AU - Palade, Vasile

PY - 2010/12/17

Y1 - 2010/12/17

N2 - This paper presents a novel algorithm for multiobjective training of Radial Basis Function (RBF) networks based on least-squares and Particle Swarm Optimization methods. The formulation is based on the fundamental concept that supervised learning is a bi-objective optimization problem, in which two conflicting objectives should be minimized. The objectives are related to the empirical training error and the machine complexity. The training is done in three steps: i) a conventional minimization of the training error, ii) multiobjective least-squares optimization for the linear parameters and, iii) particle swarm optimization for the nonlinear parameters. Some results are presented and they show the effectiveness of the proposed approach.

AB - This paper presents a novel algorithm for multiobjective training of Radial Basis Function (RBF) networks based on least-squares and Particle Swarm Optimization methods. The formulation is based on the fundamental concept that supervised learning is a bi-objective optimization problem, in which two conflicting objectives should be minimized. The objectives are related to the empirical training error and the machine complexity. The training is done in three steps: i) a conventional minimization of the training error, ii) multiobjective least-squares optimization for the linear parameters and, iii) particle swarm optimization for the nonlinear parameters. Some results are presented and they show the effectiveness of the proposed approach.

KW - Training

KW - Radial basis function networks

KW - Complexity theory

KW - Particle swarm optimization

KW - Artificial neural networks

KW - Optimization

KW - Machine learning

KW - radial basis function networks

KW - learning (artificial intelligence)

KW - least squares approximations

KW - particle swarm optimisation

KW - linear parameters

KW - multiobjective training

KW - RBF networks

KW - particle swarm optimization

KW - least-squares approximation

KW - supervised learning

KW - biobjective optimization problem

KW - machine complexity

KW - multiobjective least-squares optimization

KW - radial basis network

KW - multiobjective least squares

UR - http://www.scopus.com/inward/record.url?scp=78751565294&partnerID=8YFLogxK

U2 - 10.1109/ICTAI.2010.112

DO - 10.1109/ICTAI.2010.112

M3 - Conference proceeding

AN - SCOPUS:78751565294

SN - 978-1-4244-8817-9

VL - 2

SP - 282

EP - 285

BT - Proceedings - 22nd International Conference on Tools with Artificial Intelligence, ICTAI 2010

T2 - 22nd International Conference on Tools with Artificial Intelligence

Y2 - 27 October 2010 through 29 October 2010

ER -

On a multiobjective training algorithm for RBF networks using particle swarm optimization

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this