Orthogonal decision trees

Hillol Kargupta; Byung Hoon Park; Haimonti Dutta

doi:10.1109/TKDE.2006.127

Orthogonal decision trees

Hillol Kargupta, Byung Hoon Park, Haimonti Dutta

Performance Engineering

Research output: Contribution to journal › Article › peer-review

21 Scopus citations

Abstract

This paper introduces orthogonal decision trees that offer an effective way to construct a redundancy-free, accurate, and meaningful representation of large decision-tree-ensembles often created by popular techniques such as Bagging, Boosting, Random Forests, and many distributed and data stream mining algorithms. Orthogonal decision trees are functionally orthogonal to each other and they correspond to the principal components of the underlying function space. This paper offers a technique to construct such trees based on the Fourier transformation of decision trees and eigen-analysis of the ensemble in the Fourier representation. It offers experimental results to document the performance of orthogonal trees on the grounds of accuracy and model complexity.

Original language	English
Article number	1644727
Pages (from-to)	1028-1042
Number of pages	15
Journal	IEEE Transactions on Knowledge and Data Engineering
Volume	18
Issue number	8
DOIs	https://doi.org/10.1109/TKDE.2006.127
State	Published - Aug 2006

Funding

The authors acknowledge supports from the US National Science Foundation CAREER award IIS-0093353, US National Science Foundation grant IIS-0203958, and NASA grant NAS2-37143. The work of B.-H. Park was partially funded by the Scientific Data Management Center (http:// sdmcenter.lbl.gov) under the Department of Energy’s Scientific Discovery through Advanced Computing (DOE SciDAC) program (http://www.scidac.org). H. Kargupta is also affiliated to Agnik, LLC., Columbia, MD. A four-page version of this paper was published in the Proceedings of the 2004 IEEE International Conference on Data Mining.

Funders	Funder number
Scientific Data Management Center
US National Science Foundation	IIS-0203958
National Science Foundation	IIS-0093353
U.S. Department of Energy
National Aeronautics and Space Administration	NAS2-37143

Keywords

Fourier transform
Orthogonal decision trees
Principle component analysis
Redundancy free trees

Access to Document

10.1109/TKDE.2006.127

Cite this

@article{e20f0b1dd2514590ad4113756e88554d,

title = "Orthogonal decision trees",

abstract = "This paper introduces orthogonal decision trees that offer an effective way to construct a redundancy-free, accurate, and meaningful representation of large decision-tree-ensembles often created by popular techniques such as Bagging, Boosting, Random Forests, and many distributed and data stream mining algorithms. Orthogonal decision trees are functionally orthogonal to each other and they correspond to the principal components of the underlying function space. This paper offers a technique to construct such trees based on the Fourier transformation of decision trees and eigen-analysis of the ensemble in the Fourier representation. It offers experimental results to document the performance of orthogonal trees on the grounds of accuracy and model complexity.",

keywords = "Fourier transform, Orthogonal decision trees, Principle component analysis, Redundancy free trees",

author = "Hillol Kargupta and Park, {Byung Hoon} and Haimonti Dutta",

year = "2006",

month = aug,

doi = "10.1109/TKDE.2006.127",

language = "English",

volume = "18",

pages = "1028--1042",

journal = "IEEE Transactions on Knowledge and Data Engineering",

issn = "1041-4347",

publisher = "Institute of Electrical and Electronics Engineers",

number = "8",

}

TY - JOUR

T1 - Orthogonal decision trees

AU - Kargupta, Hillol

AU - Park, Byung Hoon

AU - Dutta, Haimonti

PY - 2006/8

Y1 - 2006/8

N2 - This paper introduces orthogonal decision trees that offer an effective way to construct a redundancy-free, accurate, and meaningful representation of large decision-tree-ensembles often created by popular techniques such as Bagging, Boosting, Random Forests, and many distributed and data stream mining algorithms. Orthogonal decision trees are functionally orthogonal to each other and they correspond to the principal components of the underlying function space. This paper offers a technique to construct such trees based on the Fourier transformation of decision trees and eigen-analysis of the ensemble in the Fourier representation. It offers experimental results to document the performance of orthogonal trees on the grounds of accuracy and model complexity.

AB - This paper introduces orthogonal decision trees that offer an effective way to construct a redundancy-free, accurate, and meaningful representation of large decision-tree-ensembles often created by popular techniques such as Bagging, Boosting, Random Forests, and many distributed and data stream mining algorithms. Orthogonal decision trees are functionally orthogonal to each other and they correspond to the principal components of the underlying function space. This paper offers a technique to construct such trees based on the Fourier transformation of decision trees and eigen-analysis of the ensemble in the Fourier representation. It offers experimental results to document the performance of orthogonal trees on the grounds of accuracy and model complexity.

KW - Fourier transform

KW - Orthogonal decision trees

KW - Principle component analysis

KW - Redundancy free trees

UR - http://www.scopus.com/inward/record.url?scp=33746359661&partnerID=8YFLogxK

U2 - 10.1109/TKDE.2006.127

DO - 10.1109/TKDE.2006.127

M3 - Article

AN - SCOPUS:33746359661

SN - 1041-4347

VL - 18

SP - 1028

EP - 1042

JO - IEEE Transactions on Knowledge and Data Engineering

JF - IEEE Transactions on Knowledge and Data Engineering

IS - 8

M1 - 1644727

ER -

Orthogonal decision trees

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this