INVESTIGADORES
GRANITTO Pablo Miguel
artículos
Título:
Feature selection on wide multiclass problems using OVA-RFE
Autor/es:
P. M. GRANITTO; A. BURGOS
Revista:
INTELIGENCIA ARTIFICIAL. IBERO-AMERICAN JOURNAL OF ARTIFICIAL INTELLIGENCE
Editorial:
AEPIA
Referencias:
Año: 2009 vol. 13 p. 27 - 34
ISSN:
1137-3601
Resumen:
Feature selection is a preprocessing technique commonly used with highdimensional datasets. It is aimed at reducing the dimensionality of the input space, discarding useless or redundant variables, in order to increase the performance and interpretability of models. For multiclass classification problems, recent works suggested that decomposing the multiclass problem in a set of binary ones, and doing the feature selection on the binary problems could be a sound strategy. In this work we combined the wellknown Recursive Feature Elimination (RFE) algorithm with the simple OneVsAll (OVA) technique for multiclass problems, to produce the new OVARFE selection method. We evaluated OVARFE using wide datasets from genomic and mass-spectrometry analysis, and several classifiers. In particular, we compared the new method with the traditional RFE (applied to a direct multiclass classifier) in terms of accuracy and stability. Our results show that OVARFE is no better than the traditional method, which is in opposition to previous results on similar methods. The opposite results are related to a diferent interpretation of the real number of variables in use by both methods.