UNIGE document Scientific Article
previous document  unige:22896  next document
add to browser collection

Variable selection in additive models by non-negative garrote

Flemming, Johanna Mills
Published in Statistical modelling. 2011, vol. 11, no. 3, p. 237-252
Abstract We adapt Breiman's non-negative garrote method to perform variable selection in non-parametric additive models. The technique avoids methods of testing for which no general reliable distributional theory is available. In addition, it removes the need for a full search of all possible models, something which is computationally intensive, especially when the number of variables is moderate to high. The method has the advantages of being conceptually simple and computationally fast. It provides accurate predictions and is effective at identifying the variables generating the model. To illustrate our procedure, we analyse logbook data on blue sharks (Prionace glauca) from the US pelagic longline fishery. In addition, we compare our proposal to a series of available alternatives by simulation. The results show that in all cases our methods perform better or as well as these alternatives.
Keywords Blue shark logbook dataCross-validationNon-negative garroteNon-parametric regressionShrinkage methodsVariable selection
Full text
Article (Published version) (174 Kb) - private document Private access
(ISO format)
CANTONI, Eva, FLEMMING, Johanna Mills, RONCHETTI, Elvezio. Variable selection in additive models by non-negative garrote. In: Statistical modelling, 2011, vol. 11, n° 3, p. 237-252. doi: 10.1177/1471082X1001100304 https://archive-ouverte.unige.ch/unige:22896

597 hits

0 download


Deposited on : 2012-09-12

Export document
Format :
Citation style :