Feature transformation and multivariate decision tree induction

Huan Liu, Rudy Setiono

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

7 Citations (Scopus)

Abstract

Univariate decision trees (UDT’s) have inherent problems of replication, repetition, and fragmentation. Multivariate decision trees (MDT’s) have been proposed to overcome some of the problems. Close examination of the conventional ways of building MDT’s, however, reveals that the fragmentation problem still persists. A novel approach is suggested to minimize the fragmentation problem by separating hyperplane search from decision tree building. This is achieved by feature transformation. Let the initial feature vector be x, the new feature vector after feature transformation T is y, i.e., y = T(x). We can obtain an MDTb y (1) building a UDT on y; and (2) replacing new features y at each node with the combinations of initial features x. We elaborate on the advantages of this approach, the details of T, and why it is expected to perform well. Experiments are conducted in order to confirm the analysis, and results are compared to those of C4.5, OC1, and CART.

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
Pages279-291
Number of pages13
Volume1532
ISBN (Print)3540653902, 9783540653905
StatePublished - 1998
Externally publishedYes
Event1st International Conference on Discovery Science, DS 1998 - Fukuoka, Japan
Duration: Dec 14 1998Dec 16 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1532
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other1st International Conference on Discovery Science, DS 1998
CountryJapan
CityFukuoka
Period12/14/9812/16/98

Fingerprint

Decision trees
Decision tree
Proof by induction
Fragmentation
Feature Vector
Univariate
Hyperplane
Replication
Minimise
Vertex of a graph
Experiment
Experiments

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Liu, H., & Setiono, R. (1998). Feature transformation and multivariate decision tree induction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1532, pp. 279-291). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1532). Springer Verlag.

Feature transformation and multivariate decision tree induction. / Liu, Huan; Setiono, Rudy.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1532 Springer Verlag, 1998. p. 279-291 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1532).

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

Liu, H & Setiono, R 1998, Feature transformation and multivariate decision tree induction. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1532, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1532, Springer Verlag, pp. 279-291, 1st International Conference on Discovery Science, DS 1998, Fukuoka, Japan, 12/14/98.
Liu H, Setiono R. Feature transformation and multivariate decision tree induction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1532. Springer Verlag. 1998. p. 279-291. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Liu, Huan ; Setiono, Rudy. / Feature transformation and multivariate decision tree induction. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1532 Springer Verlag, 1998. pp. 279-291 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{6d037a9569f24ab29f6734742bed1d35,
title = "Feature transformation and multivariate decision tree induction",
abstract = "Univariate decision trees (UDT’s) have inherent problems of replication, repetition, and fragmentation. Multivariate decision trees (MDT’s) have been proposed to overcome some of the problems. Close examination of the conventional ways of building MDT’s, however, reveals that the fragmentation problem still persists. A novel approach is suggested to minimize the fragmentation problem by separating hyperplane search from decision tree building. This is achieved by feature transformation. Let the initial feature vector be x, the new feature vector after feature transformation T is y, i.e., y = T(x). We can obtain an MDTb y (1) building a UDT on y; and (2) replacing new features y at each node with the combinations of initial features x. We elaborate on the advantages of this approach, the details of T, and why it is expected to perform well. Experiments are conducted in order to confirm the analysis, and results are compared to those of C4.5, OC1, and CART.",
author = "Huan Liu and Rudy Setiono",
year = "1998",
language = "English (US)",
isbn = "3540653902",
volume = "1532",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "279--291",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Feature transformation and multivariate decision tree induction

AU - Liu, Huan

AU - Setiono, Rudy

PY - 1998

Y1 - 1998

N2 - Univariate decision trees (UDT’s) have inherent problems of replication, repetition, and fragmentation. Multivariate decision trees (MDT’s) have been proposed to overcome some of the problems. Close examination of the conventional ways of building MDT’s, however, reveals that the fragmentation problem still persists. A novel approach is suggested to minimize the fragmentation problem by separating hyperplane search from decision tree building. This is achieved by feature transformation. Let the initial feature vector be x, the new feature vector after feature transformation T is y, i.e., y = T(x). We can obtain an MDTb y (1) building a UDT on y; and (2) replacing new features y at each node with the combinations of initial features x. We elaborate on the advantages of this approach, the details of T, and why it is expected to perform well. Experiments are conducted in order to confirm the analysis, and results are compared to those of C4.5, OC1, and CART.

AB - Univariate decision trees (UDT’s) have inherent problems of replication, repetition, and fragmentation. Multivariate decision trees (MDT’s) have been proposed to overcome some of the problems. Close examination of the conventional ways of building MDT’s, however, reveals that the fragmentation problem still persists. A novel approach is suggested to minimize the fragmentation problem by separating hyperplane search from decision tree building. This is achieved by feature transformation. Let the initial feature vector be x, the new feature vector after feature transformation T is y, i.e., y = T(x). We can obtain an MDTb y (1) building a UDT on y; and (2) replacing new features y at each node with the combinations of initial features x. We elaborate on the advantages of this approach, the details of T, and why it is expected to perform well. Experiments are conducted in order to confirm the analysis, and results are compared to those of C4.5, OC1, and CART.

UR - http://www.scopus.com/inward/record.url?scp=84937433198&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937433198&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84937433198

SN - 3540653902

SN - 9783540653905

VL - 1532

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

SP - 279

EP - 291

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

PB - Springer Verlag

ER -