Generalized Radial Basis Function networks trained with instance based learning for data mining of symbolic data

Stergios Papadimitriou, Seferina Mavroudi, Liviu Vladutu, Anastasios Bezerianos

Research output: Contribution to journalArticle

Abstract

The application of the Radial Basis Function neural networks in domains involving prediction and classification of symbolic data requires a reconsideration and a careful definition of the concept of distance between patterns. This distance in addition to providing information about the proximity of patterns should also obey some mathematical criteria in order to be applicable. Traditional distances are inadequate to access the differences between symbolic patterns. This work proposes the utilization of a statistically extracted distance measure for Generalized Radial Basis Function (GRBF) networks. The main properties of these networks are retained in the new metric space. Especially, their regularization potential can be realized with this type of distance. However, the examples of the training set for applications involving symbolic patterns are not all of the same importance and reliability. Therefore, the construction of effective decision boundaries should consider the numerous exceptions to the general motifs of classification that are frequently encountered in data mining applications. The paper supports that heuristic Instance Based Learning (IBL) training approaches can uncover information within the uneven structure of the training set. This information is exploited for the estimation of an adequate subset of the training patterns serving as RBF centers and for the estimation of effective parameter settings for those centers. The IBL learning steps are applicable to both the traditional and the statistical distance metric spaces and improve significantly the performance in both cases. The obtained results with this two-level learning method are significantly better than the traditional nearest neighbour schemes in many data mining problems.

Original languageEnglish (US)
Pages (from-to)223-234
Number of pages12
JournalApplied Intelligence
Volume16
Issue number3
DOIs
StatePublished - May 2002
Externally publishedYes

Fingerprint

Radial basis function networks
Data mining
Set theory
Neural networks

Keywords

  • Data mining
  • Heuristic learning
  • Neural network learning
  • Radial basis functions
  • Symbolic data classification

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Artificial Intelligence

Cite this

Generalized Radial Basis Function networks trained with instance based learning for data mining of symbolic data. / Papadimitriou, Stergios; Mavroudi, Seferina; Vladutu, Liviu; Bezerianos, Anastasios.

In: Applied Intelligence, Vol. 16, No. 3, 05.2002, p. 223-234.

Research output: Contribution to journalArticle

Papadimitriou, Stergios ; Mavroudi, Seferina ; Vladutu, Liviu ; Bezerianos, Anastasios. / Generalized Radial Basis Function networks trained with instance based learning for data mining of symbolic data. In: Applied Intelligence. 2002 ; Vol. 16, No. 3. pp. 223-234.
@article{761145ccfe904ef9b0af089afa735ec4,
title = "Generalized Radial Basis Function networks trained with instance based learning for data mining of symbolic data",
abstract = "The application of the Radial Basis Function neural networks in domains involving prediction and classification of symbolic data requires a reconsideration and a careful definition of the concept of distance between patterns. This distance in addition to providing information about the proximity of patterns should also obey some mathematical criteria in order to be applicable. Traditional distances are inadequate to access the differences between symbolic patterns. This work proposes the utilization of a statistically extracted distance measure for Generalized Radial Basis Function (GRBF) networks. The main properties of these networks are retained in the new metric space. Especially, their regularization potential can be realized with this type of distance. However, the examples of the training set for applications involving symbolic patterns are not all of the same importance and reliability. Therefore, the construction of effective decision boundaries should consider the numerous exceptions to the general motifs of classification that are frequently encountered in data mining applications. The paper supports that heuristic Instance Based Learning (IBL) training approaches can uncover information within the uneven structure of the training set. This information is exploited for the estimation of an adequate subset of the training patterns serving as RBF centers and for the estimation of effective parameter settings for those centers. The IBL learning steps are applicable to both the traditional and the statistical distance metric spaces and improve significantly the performance in both cases. The obtained results with this two-level learning method are significantly better than the traditional nearest neighbour schemes in many data mining problems.",
keywords = "Data mining, Heuristic learning, Neural network learning, Radial basis functions, Symbolic data classification",
author = "Stergios Papadimitriou and Seferina Mavroudi and Liviu Vladutu and Anastasios Bezerianos",
year = "2002",
month = "5",
doi = "10.1023/A:1014390017000",
language = "English (US)",
volume = "16",
pages = "223--234",
journal = "Applied Intelligence",
issn = "0924-669X",
publisher = "Springer Netherlands",
number = "3",

}

TY - JOUR

T1 - Generalized Radial Basis Function networks trained with instance based learning for data mining of symbolic data

AU - Papadimitriou, Stergios

AU - Mavroudi, Seferina

AU - Vladutu, Liviu

AU - Bezerianos, Anastasios

PY - 2002/5

Y1 - 2002/5

N2 - The application of the Radial Basis Function neural networks in domains involving prediction and classification of symbolic data requires a reconsideration and a careful definition of the concept of distance between patterns. This distance in addition to providing information about the proximity of patterns should also obey some mathematical criteria in order to be applicable. Traditional distances are inadequate to access the differences between symbolic patterns. This work proposes the utilization of a statistically extracted distance measure for Generalized Radial Basis Function (GRBF) networks. The main properties of these networks are retained in the new metric space. Especially, their regularization potential can be realized with this type of distance. However, the examples of the training set for applications involving symbolic patterns are not all of the same importance and reliability. Therefore, the construction of effective decision boundaries should consider the numerous exceptions to the general motifs of classification that are frequently encountered in data mining applications. The paper supports that heuristic Instance Based Learning (IBL) training approaches can uncover information within the uneven structure of the training set. This information is exploited for the estimation of an adequate subset of the training patterns serving as RBF centers and for the estimation of effective parameter settings for those centers. The IBL learning steps are applicable to both the traditional and the statistical distance metric spaces and improve significantly the performance in both cases. The obtained results with this two-level learning method are significantly better than the traditional nearest neighbour schemes in many data mining problems.

AB - The application of the Radial Basis Function neural networks in domains involving prediction and classification of symbolic data requires a reconsideration and a careful definition of the concept of distance between patterns. This distance in addition to providing information about the proximity of patterns should also obey some mathematical criteria in order to be applicable. Traditional distances are inadequate to access the differences between symbolic patterns. This work proposes the utilization of a statistically extracted distance measure for Generalized Radial Basis Function (GRBF) networks. The main properties of these networks are retained in the new metric space. Especially, their regularization potential can be realized with this type of distance. However, the examples of the training set for applications involving symbolic patterns are not all of the same importance and reliability. Therefore, the construction of effective decision boundaries should consider the numerous exceptions to the general motifs of classification that are frequently encountered in data mining applications. The paper supports that heuristic Instance Based Learning (IBL) training approaches can uncover information within the uneven structure of the training set. This information is exploited for the estimation of an adequate subset of the training patterns serving as RBF centers and for the estimation of effective parameter settings for those centers. The IBL learning steps are applicable to both the traditional and the statistical distance metric spaces and improve significantly the performance in both cases. The obtained results with this two-level learning method are significantly better than the traditional nearest neighbour schemes in many data mining problems.

KW - Data mining

KW - Heuristic learning

KW - Neural network learning

KW - Radial basis functions

KW - Symbolic data classification

UR - http://www.scopus.com/inward/record.url?scp=0036567394&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036567394&partnerID=8YFLogxK

U2 - 10.1023/A:1014390017000

DO - 10.1023/A:1014390017000

M3 - Article

VL - 16

SP - 223

EP - 234

JO - Applied Intelligence

JF - Applied Intelligence

SN - 0924-669X

IS - 3

ER -