Hyper-heuristic decision tree induction

Alan Vella; David Corne; Chris Murphy

doi:10.1109/NABIC.2009.5393568

Hyper-heuristic decision tree induction

Alan Vella, David Corne, Chris Murphy

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

15 Citations (Scopus)

Abstract

Hyper-heuristics are increasingly used in function and combinatorial optimization. Rather than attempt to solve a problem using a fixed heuristic, a hyper-heuristic approach attempts to find a combination of heuristics that solve a problem (and in turn may be directly suitable for a class of problem instances). Hyper-heuristics have been little explored in data mining. Here we apply a hyper-heuristic approach to data mining, by searching a space of decision tree induction algorithms. The result of hyper-heuristic search in this case is a new decision tree induction algorithm. We show that hyperheuristic search over a space of decision tree induction algorithms can find decision tree induction algorithms that outperform many different version of ID3 on unseen test sets. ©2009 IEEE.

Original language	English
Title of host publication	2009 World Congress on Nature and Biologically Inspired Computing, NABIC 2009 - Proceedings
Pages	409-414
Number of pages	6
DOIs	https://doi.org/10.1109/NABIC.2009.5393568
Publication status	Published - 2009
Event	2009 World Congress on Nature and Biologically Inspired Computing - Coimbatore, India Duration: 9 Dec 2009 → 11 Dec 2009

Conference

Conference	2009 World Congress on Nature and Biologically Inspired Computing
Abbreviated title	NABIC 2009
Country/Territory	India
City	Coimbatore
Period	9/12/09 → 11/12/09

Keywords

Data mining
Decision trees
Evolutionary algorithm
Hyper-heuristics

Access to Document

10.1109/NABIC.2009.5393568

Cite this

@inproceedings{89e8c3397f084d8eb4f7c00ce6708cea,

title = "Hyper-heuristic decision tree induction",

abstract = "Hyper-heuristics are increasingly used in function and combinatorial optimization. Rather than attempt to solve a problem using a fixed heuristic, a hyper-heuristic approach attempts to find a combination of heuristics that solve a problem (and in turn may be directly suitable for a class of problem instances). Hyper-heuristics have been little explored in data mining. Here we apply a hyper-heuristic approach to data mining, by searching a space of decision tree induction algorithms. The result of hyper-heuristic search in this case is a new decision tree induction algorithm. We show that hyperheuristic search over a space of decision tree induction algorithms can find decision tree induction algorithms that outperform many different version of ID3 on unseen test sets. {\textcopyright}2009 IEEE.",

keywords = "Data mining, Decision trees, Evolutionary algorithm, Hyper-heuristics",

author = "Alan Vella and David Corne and Chris Murphy",

year = "2009",

doi = "10.1109/NABIC.2009.5393568",

language = "English",

isbn = "9781424456123",

pages = "409--414",

booktitle = "2009 World Congress on Nature and Biologically Inspired Computing, NABIC 2009 - Proceedings",

note = "2009 World Congress on Nature and Biologically Inspired Computing, NABIC 2009 ; Conference date: 09-12-2009 Through 11-12-2009",

}

TY - GEN

T1 - Hyper-heuristic decision tree induction

AU - Vella, Alan

AU - Corne, David

AU - Murphy, Chris

PY - 2009

Y1 - 2009

N2 - Hyper-heuristics are increasingly used in function and combinatorial optimization. Rather than attempt to solve a problem using a fixed heuristic, a hyper-heuristic approach attempts to find a combination of heuristics that solve a problem (and in turn may be directly suitable for a class of problem instances). Hyper-heuristics have been little explored in data mining. Here we apply a hyper-heuristic approach to data mining, by searching a space of decision tree induction algorithms. The result of hyper-heuristic search in this case is a new decision tree induction algorithm. We show that hyperheuristic search over a space of decision tree induction algorithms can find decision tree induction algorithms that outperform many different version of ID3 on unseen test sets. ©2009 IEEE.

AB - Hyper-heuristics are increasingly used in function and combinatorial optimization. Rather than attempt to solve a problem using a fixed heuristic, a hyper-heuristic approach attempts to find a combination of heuristics that solve a problem (and in turn may be directly suitable for a class of problem instances). Hyper-heuristics have been little explored in data mining. Here we apply a hyper-heuristic approach to data mining, by searching a space of decision tree induction algorithms. The result of hyper-heuristic search in this case is a new decision tree induction algorithm. We show that hyperheuristic search over a space of decision tree induction algorithms can find decision tree induction algorithms that outperform many different version of ID3 on unseen test sets. ©2009 IEEE.

KW - Data mining

KW - Decision trees

KW - Evolutionary algorithm

KW - Hyper-heuristics

UR - http://www.scopus.com/inward/record.url?scp=77949578105&partnerID=8YFLogxK

U2 - 10.1109/NABIC.2009.5393568

DO - 10.1109/NABIC.2009.5393568

M3 - Conference contribution

SN - 9781424456123

SP - 409

EP - 414

BT - 2009 World Congress on Nature and Biologically Inspired Computing, NABIC 2009 - Proceedings

T2 - 2009 World Congress on Nature and Biologically Inspired Computing

Y2 - 9 December 2009 through 11 December 2009

ER -

Hyper-heuristic decision tree induction

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this