GA-Stacking: Evolutionary Stacked Generalization. Ledezma, A., Aler, R., Sanchis, A., & Borrajo, D. Intelligent Data Analysis, IOS Press, 2009. Accepted
abstract   bibtex   
Stacking is one of the most used techniques for combining classifiers and improve prediction accuracy. Early research in stacking showed that selecting the right classifiers, their parameters and the metaclassifiers was the main bottleneck for its use. Most of the research on this topic selects by hand the right combination of classifiers and their parameters. Instead of starting from these initial strong assumptions, our approach uses genetic algorithms to search for good stacking configurations. Since this can lead to overfitting, one of the goals of this paper is to evaluate empirically the overall efficiency of the approach. A second goal is to compare our approach with current best stacking building techniques. The results show that our approach finds stacking configurations that, in the worst case, perform as well as the best techniques, with the advantage of not having to set up manually the structure of the stacking system.
@ARTICLE{jida09-gastacking,
  author = {Agapito Ledezma and Ricardo Aler and Araceli Sanchis and Daniel Borrajo},
  title = {GA-Stacking: Evolutionary Stacked Generalization},
  journal = {Intelligent Data Analysis},
  year = {2009},
  volume = {14},
  note = {Accepted},
  abstract = {Stacking is one of the most used techniques for combining classifiers
	and improve prediction accuracy. Early research in stacking showed
	that selecting the right classifiers, their parameters and the metaclassifiers
	was the main bottleneck for its use. Most of the research on this
	topic selects by hand the right combination of classifiers and their
	parameters. Instead of starting from these initial strong assumptions,
	our approach uses genetic algorithms to search for good stacking
	configurations. Since this can lead to overfitting, one of the goals
	of this paper is to evaluate empirically the overall efficiency of
	the approach. A second goal is to compare our approach with current
	best stacking building techniques. The
	
	results show that our approach finds stacking configurations that,
	in the worst case, perform as well as the best techniques, with the
	advantage of not having to set up manually the structure of the stacking
	system.},
  bib2html_pubtype = {Journal},
  bib2html_rescat = {Ensemble of classifiers},
  country = {Netherlands},
  issn = {1088-467X (Print) 1571-4128 (Online)},
  jcr = {2007: 0.446 (76/93), 2008: 0.426 (86/94)},
  jpublisher = {IOS Press},
  owner = {ledezma},
  publisher = {IOS Press},
  timestamp = {2011.11.21}
}

Downloads: 0