Publications:Leverages Based Neural Networks Fusion

Do not edit this section

Property "Author" has a restricted application area and cannot be used as annotation property by a user. Property "Author" has a restricted application area and cannot be used as annotation property by a user. Property "Author" has a restricted application area and cannot be used as annotation property by a user.

Keep all hand-made modifications below

Title	Leverages Based Neural Networks Fusion
Author
Year	2004
PublicationType	Conference Paper
Journal
HostPublication	Neural information processing
Conference	11th International Conference, ICONIP 2004, Calcutta
DOI	http://dx.doi.org/10.1007/978-3-540-30499-9_68
Diva url	http://hh.diva-portal.org/smash/record.jsf?searchId=1&pid=diva2:300225
Abstract	To improve estimation results, outputs of multiple neural networks can be aggregated into a committee output. In this paper, we study the usefulness of the leverages based information for creating accurate neural network committees. Based on the approximate leave-one-out error and the suggested, generalization error based, diversity test, accurate and diverse networks are selected and fused into a committee using data dependent aggregation weights. Four data dependent aggregation schemes – based on local variance, covariance, Choquet integral, and the generalized Choquet integral – are investigated. The effectiveness of the approaches is tested on one artificial and three real world data sets.