TY - GEN
T1 - A new landmarker generation algorithm based on correlativity
AU - Ler, Daren
AU - Koprinska, Irena
AU - Chawla, Sanjay
PY - 2004
Y1 - 2004
N2 - Landmarking is a recent and promising metalearning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. With these, we introduce a landmarker generation algorithm, which creates a set of landmarkers that each utilise subsets of the algorithms being landmarked. The experiments show that the landmarkers formed, when used with linear regression, are able to estimate accuracy well, even when utilising a small fraction of the given algorithms.
AB - Landmarking is a recent and promising metalearning strategy, which defines meta-features that are themselves efficient learning algorithms. However, the choice of landmarkers is made in an ad hoc manner. In this paper, we propose a new perspective and set of criteria for landmarkers. With these, we introduce a landmarker generation algorithm, which creates a set of landmarkers that each utilise subsets of the algorithms being landmarked. The experiments show that the landmarkers formed, when used with linear regression, are able to estimate accuracy well, even when utilising a small fraction of the given algorithms.
UR - http://www.scopus.com/inward/record.url?scp=21244449557&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:21244449557
SN - 0780388232
T3 - Proceedings of the 2004 International Conference on Machine Learning and Applications, ICMLA '04
SP - 178
EP - 185
BT - Proceedings of the 2004 International Conference on Machine Learning and Applications, ICMLA '04
A2 - Kantardzic, M.
A2 - Nasraoui, O.
A2 - Milanova, M.
T2 - 2004 International Conference on Machine Learning and Applications, ICMLA '04
Y2 - 16 December 2004 through 18 December 2004
ER -