Neurolinear: A system for extracting oblique decision rules from neural networks

Rudy Setiono, Huan Liu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

We present NeuroLinear, a system for extracting oblique decision rules from neural networks that have been trained for classification of patterns. Each condition of an oblique decision rule corresponds to a partition of the attribute space by a hyperplane that is not necessarily axis-parallel. Allowing a set of such hyperplanes to form the boundaries of the decision regions leads to a significant reduction in the number of rules generated while maintaining the accuracy rates of the networks. We describe the components of NeuroLinear in detail using a heart disease diagnosis problem. Our experimental results on real-world datasets show that the system is effective in extracting compact and comprehensible rules with high predictive accuracy from neural networks.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages221-233
Number of pages13
Volume1224
ISBN (Print)3540628584, 9783540628583
StatePublished - 1997
Externally publishedYes
Event9th European Conference on Machine Learning, ECML 1997 - Prague, Czech Republic
Duration: Apr 23 1997Apr 25 1997

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1224
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other9th European Conference on Machine Learning, ECML 1997
CountryCzech Republic
CityPrague
Period4/23/974/25/97

Fingerprint

Decision Rules
Oblique
Hyperplane
Neural Networks
Neural networks
Attribute
Partition
Experimental Results
Heart

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Setiono, R., & Liu, H. (1997). Neurolinear: A system for extracting oblique decision rules from neural networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1224, pp. 221-233). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1224). Springer Verlag.

Neurolinear : A system for extracting oblique decision rules from neural networks. / Setiono, Rudy; Liu, Huan.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1224 Springer Verlag, 1997. p. 221-233 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1224).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Setiono, R & Liu, H 1997, Neurolinear: A system for extracting oblique decision rules from neural networks. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1224, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1224, Springer Verlag, pp. 221-233, 9th European Conference on Machine Learning, ECML 1997, Prague, Czech Republic, 4/23/97.
Setiono R, Liu H. Neurolinear: A system for extracting oblique decision rules from neural networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1224. Springer Verlag. 1997. p. 221-233. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Setiono, Rudy ; Liu, Huan. / Neurolinear : A system for extracting oblique decision rules from neural networks. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1224 Springer Verlag, 1997. pp. 221-233 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{039d264170954d8cb2ef410bffa6272c,
title = "Neurolinear: A system for extracting oblique decision rules from neural networks",
abstract = "We present NeuroLinear, a system for extracting oblique decision rules from neural networks that have been trained for classification of patterns. Each condition of an oblique decision rule corresponds to a partition of the attribute space by a hyperplane that is not necessarily axis-parallel. Allowing a set of such hyperplanes to form the boundaries of the decision regions leads to a significant reduction in the number of rules generated while maintaining the accuracy rates of the networks. We describe the components of NeuroLinear in detail using a heart disease diagnosis problem. Our experimental results on real-world datasets show that the system is effective in extracting compact and comprehensible rules with high predictive accuracy from neural networks.",
author = "Rudy Setiono and Huan Liu",
year = "1997",
language = "English (US)",
isbn = "3540628584",
volume = "1224",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "221--233",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Neurolinear

T2 - A system for extracting oblique decision rules from neural networks

AU - Setiono, Rudy

AU - Liu, Huan

PY - 1997

Y1 - 1997

N2 - We present NeuroLinear, a system for extracting oblique decision rules from neural networks that have been trained for classification of patterns. Each condition of an oblique decision rule corresponds to a partition of the attribute space by a hyperplane that is not necessarily axis-parallel. Allowing a set of such hyperplanes to form the boundaries of the decision regions leads to a significant reduction in the number of rules generated while maintaining the accuracy rates of the networks. We describe the components of NeuroLinear in detail using a heart disease diagnosis problem. Our experimental results on real-world datasets show that the system is effective in extracting compact and comprehensible rules with high predictive accuracy from neural networks.

AB - We present NeuroLinear, a system for extracting oblique decision rules from neural networks that have been trained for classification of patterns. Each condition of an oblique decision rule corresponds to a partition of the attribute space by a hyperplane that is not necessarily axis-parallel. Allowing a set of such hyperplanes to form the boundaries of the decision regions leads to a significant reduction in the number of rules generated while maintaining the accuracy rates of the networks. We describe the components of NeuroLinear in detail using a heart disease diagnosis problem. Our experimental results on real-world datasets show that the system is effective in extracting compact and comprehensible rules with high predictive accuracy from neural networks.

UR - http://www.scopus.com/inward/record.url?scp=70249083509&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70249083509&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:70249083509

SN - 3540628584

SN - 9783540628583

VL - 1224

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 221

EP - 233

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -