Hybrid Pruning Algorithm

Xiangran DU; Xizhao WANG; Yuanyuan WAN

doi:10.1109/IFCSTA.2009.13

Hybrid Pruning Algorithm

Xiangran DU^*, Xizhao WANG, Yuanyuan WAN

^*Corresponding author for this work

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

Abstract

In this paper we develop a new post-pruning algorithm. This new pruning algorithm uses two or more postpruning algorithms to prune a decision tree that has been built on training set by different orders, and the "best" tree is selected based either on separate test set accuracy or cross-validations from trees coming from result of the above step. The algorithm is theoretically based on occam's razor that is a simpler model is chosen if two models have the same performance on the training set. An experiment is implemented on three databases in UCI machine learning repository and the new algorithm is employed to compares with two well-known post-pruning algorithms. The results show that the hybrid pruning algorithm effectively reduces the complexity of decision trees without sacrificing accuracy.

Original language	English
Title of host publication	Proceedings : 2009 International Forum on Computer Science-Technology and Applications
Publisher	IEEE
Pages	30-33
Number of pages	4
Volume	3
ISBN (Electronic)	9781424454235
ISBN (Print)	9780769539300, 9781424454228
DOIs	https://doi.org/10.1109/IFCSTA.2009.13
Publication status	Published - 2009
Externally published	Yes
Event	2009 International Forum on Computer Science-Technology and Applications, IFCSTA 2009 - Chongqing, China Duration: 25 Dec 2009 → 27 Dec 2009

Publication series

Name	International Forum on Computer Science-Technology and Applications (IFCSTA)
Publisher	IEEE

Conference

Conference	2009 International Forum on Computer Science-Technology and Applications, IFCSTA 2009
Country/Territory	China
City	Chongqing
Period	25/12/09 → 27/12/09

Bibliographical note

This research is supported by the Natural Science Foundation of Hebei Province (F2008000635), by the key project foundation of applied fundamental research of Hebei Province (08963522D), by the plan of 100 excellent innovative scientists of the first group in Education Department of Hebei Province, and by the Scientific
Research Foundation of Hebei Province (06213548).

Keywords

Decision tree simplification
Decision trees
Hybrid pruning algorithm
Occam's razor
Overfitting

Access to Document

10.1109/IFCSTA.2009.13

Cite this

@inproceedings{4bf37e5f37a44adeb186911c6ebb114c,

title = "Hybrid Pruning Algorithm",

abstract = "In this paper we develop a new post-pruning algorithm. This new pruning algorithm uses two or more postpruning algorithms to prune a decision tree that has been built on training set by different orders, and the {"}best{"} tree is selected based either on separate test set accuracy or cross-validations from trees coming from result of the above step. The algorithm is theoretically based on occam's razor that is a simpler model is chosen if two models have the same performance on the training set. An experiment is implemented on three databases in UCI machine learning repository and the new algorithm is employed to compares with two well-known post-pruning algorithms. The results show that the hybrid pruning algorithm effectively reduces the complexity of decision trees without sacrificing accuracy.",

keywords = "Decision tree simplification, Decision trees, Hybrid pruning algorithm, Occam's razor, Overfitting",

author = "Xiangran DU and Xizhao WANG and Yuanyuan WAN",

note = "This research is supported by the Natural Science Foundation of Hebei Province (F2008000635), by the key project foundation of applied fundamental research of Hebei Province (08963522D), by the plan of 100 excellent innovative scientists of the first group in Education Department of Hebei Province, and by the Scientific Research Foundation of Hebei Province (06213548).; 2009 International Forum on Computer Science-Technology and Applications, IFCSTA 2009 ; Conference date: 25-12-2009 Through 27-12-2009",

year = "2009",

doi = "10.1109/IFCSTA.2009.13",

language = "English",

isbn = "9780769539300",

volume = "3",

series = "International Forum on Computer Science-Technology and Applications (IFCSTA)",

publisher = "IEEE",

pages = "30--33",

booktitle = "Proceedings : 2009 International Forum on Computer Science-Technology and Applications",

}

DU, X, WANG, X & WAN, Y 2009, Hybrid Pruning Algorithm. in Proceedings : 2009 International Forum on Computer Science-Technology and Applications. vol. 3, 5385140, International Forum on Computer Science-Technology and Applications (IFCSTA), IEEE, pp. 30-33, 2009 International Forum on Computer Science-Technology and Applications, IFCSTA 2009, Chongqing, China, 25/12/09. https://doi.org/10.1109/IFCSTA.2009.13

Hybrid Pruning Algorithm. / DU, Xiangran; WANG, Xizhao; WAN, Yuanyuan.
Proceedings : 2009 International Forum on Computer Science-Technology and Applications. Vol. 3 IEEE, 2009. p. 30-33 5385140 (International Forum on Computer Science-Technology and Applications (IFCSTA)).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

TY - GEN

T1 - Hybrid Pruning Algorithm

AU - DU, Xiangran

AU - WANG, Xizhao

AU - WAN, Yuanyuan

N1 - This research is supported by the Natural Science Foundation of Hebei Province (F2008000635), by the key project foundation of applied fundamental research of Hebei Province (08963522D), by the plan of 100 excellent innovative scientists of the first group in Education Department of Hebei Province, and by the Scientific Research Foundation of Hebei Province (06213548).

PY - 2009

Y1 - 2009

N2 - In this paper we develop a new post-pruning algorithm. This new pruning algorithm uses two or more postpruning algorithms to prune a decision tree that has been built on training set by different orders, and the "best" tree is selected based either on separate test set accuracy or cross-validations from trees coming from result of the above step. The algorithm is theoretically based on occam's razor that is a simpler model is chosen if two models have the same performance on the training set. An experiment is implemented on three databases in UCI machine learning repository and the new algorithm is employed to compares with two well-known post-pruning algorithms. The results show that the hybrid pruning algorithm effectively reduces the complexity of decision trees without sacrificing accuracy.

AB - In this paper we develop a new post-pruning algorithm. This new pruning algorithm uses two or more postpruning algorithms to prune a decision tree that has been built on training set by different orders, and the "best" tree is selected based either on separate test set accuracy or cross-validations from trees coming from result of the above step. The algorithm is theoretically based on occam's razor that is a simpler model is chosen if two models have the same performance on the training set. An experiment is implemented on three databases in UCI machine learning repository and the new algorithm is employed to compares with two well-known post-pruning algorithms. The results show that the hybrid pruning algorithm effectively reduces the complexity of decision trees without sacrificing accuracy.

KW - Decision tree simplification

KW - Decision trees

KW - Hybrid pruning algorithm

KW - Occam's razor

KW - Overfitting

UR - http://www.scopus.com/inward/record.url?scp=77949852080&partnerID=8YFLogxK

U2 - 10.1109/IFCSTA.2009.13

DO - 10.1109/IFCSTA.2009.13

M3 - Conference paper (refereed)

AN - SCOPUS:77949852080

SN - 9780769539300

SN - 9781424454228

VL - 3

T3 - International Forum on Computer Science-Technology and Applications (IFCSTA)

SP - 30

EP - 33

BT - Proceedings : 2009 International Forum on Computer Science-Technology and Applications

PB - IEEE

T2 - 2009 International Forum on Computer Science-Technology and Applications, IFCSTA 2009

Y2 - 25 December 2009 through 27 December 2009

ER -

Hybrid Pruning Algorithm

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this