Off-line structural risk minimization and BARTMAP-S

Stephen J. Verzi; Gregory L. Heileman; Michael Georgiopoulos; Georgios Anagnostopoulos

Off-line structural risk minimization and BARTMAP-S

Stephen J. Verzi, Gregory L. Heileman, Michael Georgiopoulos, Georgios Anagnostopoulos

Research output: Contribution to conference › Paper › peer-review

Abstract

BARTMAP-S introduced a neural network architecture with which structural risk minimization can be performed, although indirectly. BARTMAP-S as previously described is trained in an on-line fashion, consistent with the original way intended for the Fuzzy ARTMAP neural network architecture. Here we will propose an extension to BARTMAP-S for conducting off-line learning. Consequently, this alternate mode of learning will allow us to conduct structural risk minimization more directly. In this paper, we will describe the new architecture and present some empirical results to demonstrate the usefulness of structural risk minimization in learning with an ARTMAP-based neural network.

Original language	English (US)
Pages	2533-2538
Number of pages	6
State	Published - 2002
Externally published	Yes
Event	2002 International Joint Conference on Neural Networks (IJCNN'02) - Honolulu, HI, United States Duration: May 12 2002 → May 17 2002

Conference

Conference	2002 International Joint Conference on Neural Networks (IJCNN'02)
Country/Territory	United States
City	Honolulu, HI
Period	5/12/02 → 5/17/02

Keywords

Adaptive resonance theory
Classification
Empirical risk minimization
Generalization performance
Machine learning
Neural networks
Overlapping pattern classes
Structural risk minimization

ASJC Scopus subject areas

Software
Artificial Intelligence

Cite this

@conference{ebf95058f9cc4d59ac1fc9a118edf2ae,

title = "Off-line structural risk minimization and BARTMAP-S",

abstract = "BARTMAP-S introduced a neural network architecture with which structural risk minimization can be performed, although indirectly. BARTMAP-S as previously described is trained in an on-line fashion, consistent with the original way intended for the Fuzzy ARTMAP neural network architecture. Here we will propose an extension to BARTMAP-S for conducting off-line learning. Consequently, this alternate mode of learning will allow us to conduct structural risk minimization more directly. In this paper, we will describe the new architecture and present some empirical results to demonstrate the usefulness of structural risk minimization in learning with an ARTMAP-based neural network.",

keywords = "Adaptive resonance theory, Classification, Empirical risk minimization, Generalization performance, Machine learning, Neural networks, Overlapping pattern classes, Structural risk minimization",

author = "Verzi, {Stephen J.} and Heileman, {Gregory L.} and Michael Georgiopoulos and Georgios Anagnostopoulos",

year = "2002",

language = "English (US)",

pages = "2533--2538",

note = "2002 International Joint Conference on Neural Networks (IJCNN'02) ; Conference date: 12-05-2002 Through 17-05-2002",

}

TY - CONF

T1 - Off-line structural risk minimization and BARTMAP-S

AU - Verzi, Stephen J.

AU - Heileman, Gregory L.

AU - Georgiopoulos, Michael

AU - Anagnostopoulos, Georgios

PY - 2002

Y1 - 2002

N2 - BARTMAP-S introduced a neural network architecture with which structural risk minimization can be performed, although indirectly. BARTMAP-S as previously described is trained in an on-line fashion, consistent with the original way intended for the Fuzzy ARTMAP neural network architecture. Here we will propose an extension to BARTMAP-S for conducting off-line learning. Consequently, this alternate mode of learning will allow us to conduct structural risk minimization more directly. In this paper, we will describe the new architecture and present some empirical results to demonstrate the usefulness of structural risk minimization in learning with an ARTMAP-based neural network.

AB - BARTMAP-S introduced a neural network architecture with which structural risk minimization can be performed, although indirectly. BARTMAP-S as previously described is trained in an on-line fashion, consistent with the original way intended for the Fuzzy ARTMAP neural network architecture. Here we will propose an extension to BARTMAP-S for conducting off-line learning. Consequently, this alternate mode of learning will allow us to conduct structural risk minimization more directly. In this paper, we will describe the new architecture and present some empirical results to demonstrate the usefulness of structural risk minimization in learning with an ARTMAP-based neural network.

KW - Adaptive resonance theory

KW - Classification

KW - Empirical risk minimization

KW - Generalization performance

KW - Machine learning

KW - Neural networks

KW - Overlapping pattern classes

KW - Structural risk minimization

UR - http://www.scopus.com/inward/record.url?scp=0036082980&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036082980&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0036082980

SP - 2533

EP - 2538

T2 - 2002 International Joint Conference on Neural Networks (IJCNN'02)

Y2 - 12 May 2002 through 17 May 2002

ER -

Off-line structural risk minimization and BARTMAP-S

Abstract

Conference

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this