Negative correlation learning and the ambiguity family of ensemble methods

Negative correlation learning and the ambiguity family of ensemble methods (2003)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Gavin Brown , Jeremy Wyatt

Venue:	In Proc. Int. Workshop on Multiple Classifier Systems (LNCS 2709
Citations:	8 - 1 self

Versions

Version History

Metadata Version 1

Datum	Value	Source
TITLE	Negative correlation learning and the ambiguity family of ensemble methods	INFERENCE
AUTHOR NAME	Gavin Brown	SVM HeaderParse 0.2
AUTHOR AFFIL	School of Computer Science, University of Birmingham; Edgbaston Park Road, Birmingham	SVM HeaderParse 0.2
AUTHOR ADDR	B15 2TT	SVM HeaderParse 0.2
AUTHOR NAME	Jeremy Wyatt	SVM HeaderParse 0.2
AUTHOR AFFIL	School of Computer Science, University of Birmingham; Edgbaston Park Road, Birmingham	SVM HeaderParse 0.2
AUTHOR ADDR	B15 2TT	SVM HeaderParse 0.2
ABSTRACT	Abstract. We study the formal basis behind Negative Correlation (NC) Learning, an ensemble technique developed in the evolutionary computation literature. We show that by removing an assumption made in the original work, NC can be shown to be a derivative technique of the Ambiguity decomposition by Krogh and Vedelsby. From this formalisation, we calculate parameter bounds, and show significant improvements in empirical tests. We hypothesize that the reason for its success lies in rescaling an estimate of ensemble covariance; then show that during this rescaling, NC varies smoothly between a single neural network and an ensemble system. Finally we unify several other works in the literature, all of which have exploited the Ambiguity decomposition in some way, and term them the Ambiguity Family. 1	SVM HeaderParse 0.2
YEAR	2003	INFERENCE
VENUE	In Proc. Int. Workshop on Multiple Classifier Systems (LNCS 2709	INFERENCE
VENUE TYPE	CONFERENCE	INFERENCE
PAGES	266--275	INFERENCE
VOLUME	2709	INFERENCE
CITATIONS	12 found	ParsCit 1.0