Cross validation in LULOO

Paul Haase Sørensen; Peter Magnus Nørgård; Lars Kai Hansen; Jan Larsen

Cross validation in LULOO

Paul Haase Sørensen, Peter Magnus Nørgård, Lars Kai Hansen, Jan Larsen

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

The leave-one-out cross-validation scheme for generalization assessment of neural network models is computationally expensive due to replicated training sessions. Linear unlearning of examples has recently been suggested as an approach to approximative cross-validation. Here we briefly review the linear unlearning scheme, dubbed LULOO, and we illustrate it on a systemidentification example. Further, we address the possibility of extracting confidence information (error bars) from the LULOO ensemble.

Original language	English
Title of host publication	Proceedings of International Conference on Neural Information Processing
Publication date	1996
Publication status	Published - 1996
Event	International Conference on Neural Information Processing - Hong Kong Duration: 1 Jan 1996 → …

Conference

Conference	International Conference on Neural Information Processing
City	Hong Kong
Period	01/01/1996 → …

Cite this

@inproceedings{1ff2e203f8c74c98b4eee47a15d1afab,

title = "Cross validation in LULOO",

abstract = "The leave-one-out cross-validation scheme for generalization assessment of neural network models is computationally expensive due to replicated training sessions. Linear unlearning of examples has recently been suggested as an approach to approximative cross-validation. Here we briefly review the linear unlearning scheme, dubbed LULOO, and we illustrate it on a systemidentification example. Further, we address the possibility of extracting confidence information (error bars) from the LULOO ensemble.",

author = "S{\o}rensen, {Paul Haase} and N{\o}rg{\aa}rd, {Peter Magnus} and Hansen, {Lars Kai} and Jan Larsen",

year = "1996",

language = "English",

booktitle = "Proceedings of International Conference on Neural Information Processing",

note = "International Conference on Neural Information Processing ; Conference date: 01-01-1996",

}

TY - GEN

T1 - Cross validation in LULOO

AU - Sørensen, Paul Haase

AU - Nørgård, Peter Magnus

AU - Hansen, Lars Kai

AU - Larsen, Jan

PY - 1996

Y1 - 1996

N2 - The leave-one-out cross-validation scheme for generalization assessment of neural network models is computationally expensive due to replicated training sessions. Linear unlearning of examples has recently been suggested as an approach to approximative cross-validation. Here we briefly review the linear unlearning scheme, dubbed LULOO, and we illustrate it on a systemidentification example. Further, we address the possibility of extracting confidence information (error bars) from the LULOO ensemble.

AB - The leave-one-out cross-validation scheme for generalization assessment of neural network models is computationally expensive due to replicated training sessions. Linear unlearning of examples has recently been suggested as an approach to approximative cross-validation. Here we briefly review the linear unlearning scheme, dubbed LULOO, and we illustrate it on a systemidentification example. Further, we address the possibility of extracting confidence information (error bars) from the LULOO ensemble.

M3 - Article in proceedings

BT - Proceedings of International Conference on Neural Information Processing

T2 - International Conference on Neural Information Processing

Y2 - 1 January 1996

ER -