Feature transformation and multivariate decision tree induction

Huan Liu, Rudy Setiono

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

10 Scopus citations

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 publicationDiscovery Science - 1st International Conference, DS 1998, Proceedings
EditorsSetsuo Arikawa, Hiroshi Motoda
PublisherSpringer Verlag
Pages279-291
Number of pages13
ISBN (Print)3540653902, 9783540653905
DOIs
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

Other

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

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Feature transformation and multivariate decision tree induction'. Together they form a unique fingerprint.

Cite this