A landmarker selection algorithm based on correlation and efficiency criteria

Daren Ler, Irena Koprinska, Sanjay Chawla

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

1 Citation (Scopus)

Abstract

Landmarking is a recent and promising meta-learning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is often made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. Based on the new criteria, we propose a landmarker generation algorithm, which generates a set of landmarkers that are each subsets of the algorithms being landmarked. Our experiments show that the landmarkers formed, when used with linear regression are able to estimate the accuracy of a set of candidate algorithms well, while only utilising a small fraction of the computational cost required to evaluate those candidate algorithms via ten-fold cross-validation.

Original languageEnglish
Title of host publicationLecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)
EditorsG.I. Webb, X. Yu
Pages296-306
Number of pages11
Volume3339
Publication statusPublished - 2004
Externally publishedYes
Event17th Australian Joint Conference on Artificial Intelligence, AI 2004: Advances in Artificial Intelligence - Cairns, Australia
Duration: 4 Dec 20046 Dec 2004

Other

Other17th Australian Joint Conference on Artificial Intelligence, AI 2004: Advances in Artificial Intelligence
CountryAustralia
CityCairns
Period4/12/046/12/04

Fingerprint

Set theory
Linear regression
Learning algorithms
Costs
Experiments

ASJC Scopus subject areas

  • Hardware and Architecture

Cite this

Ler, D., Koprinska, I., & Chawla, S. (2004). A landmarker selection algorithm based on correlation and efficiency criteria. In G. I. Webb, & X. Yu (Eds.), Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) (Vol. 3339, pp. 296-306)

A landmarker selection algorithm based on correlation and efficiency criteria. / Ler, Daren; Koprinska, Irena; Chawla, Sanjay.

Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). ed. / G.I. Webb; X. Yu. Vol. 3339 2004. p. 296-306.

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

Ler, D, Koprinska, I & Chawla, S 2004, A landmarker selection algorithm based on correlation and efficiency criteria. in GI Webb & X Yu (eds), Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). vol. 3339, pp. 296-306, 17th Australian Joint Conference on Artificial Intelligence, AI 2004: Advances in Artificial Intelligence, Cairns, Australia, 4/12/04.
Ler D, Koprinska I, Chawla S. A landmarker selection algorithm based on correlation and efficiency criteria. In Webb GI, Yu X, editors, Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). Vol. 3339. 2004. p. 296-306
Ler, Daren ; Koprinska, Irena ; Chawla, Sanjay. / A landmarker selection algorithm based on correlation and efficiency criteria. Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). editor / G.I. Webb ; X. Yu. Vol. 3339 2004. pp. 296-306
@inproceedings{5bb7593f1d1d4513a83051ea5f891c6d,
title = "A landmarker selection algorithm based on correlation and efficiency criteria",
abstract = "Landmarking is a recent and promising meta-learning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is often made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. Based on the new criteria, we propose a landmarker generation algorithm, which generates a set of landmarkers that are each subsets of the algorithms being landmarked. Our experiments show that the landmarkers formed, when used with linear regression are able to estimate the accuracy of a set of candidate algorithms well, while only utilising a small fraction of the computational cost required to evaluate those candidate algorithms via ten-fold cross-validation.",
author = "Daren Ler and Irena Koprinska and Sanjay Chawla",
year = "2004",
language = "English",
volume = "3339",
pages = "296--306",
editor = "G.I. Webb and X. Yu",
booktitle = "Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)",

}

TY - GEN

T1 - A landmarker selection algorithm based on correlation and efficiency criteria

AU - Ler, Daren

AU - Koprinska, Irena

AU - Chawla, Sanjay

PY - 2004

Y1 - 2004

N2 - Landmarking is a recent and promising meta-learning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is often made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. Based on the new criteria, we propose a landmarker generation algorithm, which generates a set of landmarkers that are each subsets of the algorithms being landmarked. Our experiments show that the landmarkers formed, when used with linear regression are able to estimate the accuracy of a set of candidate algorithms well, while only utilising a small fraction of the computational cost required to evaluate those candidate algorithms via ten-fold cross-validation.

AB - Landmarking is a recent and promising meta-learning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is often made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. Based on the new criteria, we propose a landmarker generation algorithm, which generates a set of landmarkers that are each subsets of the algorithms being landmarked. Our experiments show that the landmarkers formed, when used with linear regression are able to estimate the accuracy of a set of candidate algorithms well, while only utilising a small fraction of the computational cost required to evaluate those candidate algorithms via ten-fold cross-validation.

UR - http://www.scopus.com/inward/record.url?scp=22944456185&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=22944456185&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:22944456185

VL - 3339

SP - 296

EP - 306

BT - Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)

A2 - Webb, G.I.

A2 - Yu, X.

ER -