Statistical Modelling 11 (2011), 237–252

Variable selection in additive models by non-negative garrote

Eva Cantoni
Institute of Statistics and Department of Economics,
University of Geneva
CH–1211 Geneva
Switzerland
eMail: eva.cantoni@unige.ch

Joanna Mills Flemming
Department of Mathematics and Statistics,
Dalhousie University
Canada

Elvezio Ronchetti
Institute of Statistics and Department of Economics,
University of Geneva
Switzerland

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 data; cross-validation; non-negative garrote; non-parametric regression; shrinkage methods; variable selection

Downloads:

Example data and code in zipped archive
back