Meta-spv learning

Concepts

These components integrate supervised learning algorithms. They can produce one or more instances of the supervised learning algorithm, and make them to cooperate.

These components must be used with "supervised learning" component.

All kind of supervised learning algorithm can be combined.

Main survey about these methods is
E. Bauer, R. Kohavi, "An empirical comparison of voting classification algorithms : bagging, boosting and variants", Machine Learning, vol. 36, pp.105-139, 1999.

Attributes status

See supervised learning components.

Meta-spv learning components

Component Description Parameters Note

Instance Spv Learning (Supervised Learning)
Instanciate a single supervised learning algorithm. Basic meta-spv learning

Arcing [Arc-x4]
Run several learning process, with reweighted examples.

- L. Breiman, "Arcing classifiers", The Annals of Statistics, vol. 26, pp.801-849, 1998.

Number of classifiers.

Boosting
Similar to Arcing.

- Y. Freund, R.E. Schapire, "Experiments with a new boosting algorithm", Proceedings of 13th ICML, pp. 148-156, 1996.

Number of classifiers.

Bagging
Run several learning process, examples are sampled with replication.

- L. Breiman, "Bagging predictors", Machine Learning, 24(2), pp. 123-140, 1996.

Number of classifiers.

Last modification : January 21st, 2004.