Variable selection through CART

Marie Sauvé (1) , Christine Tuleau (1)

1 SELECT - Model selection in statistical learning

This paper deals with variable selection in the regression or binary classification frameworks. It proposes an automatic and exhaustive procedure which relies on the use of the CART algorithm and on model selection via penalization. This work, of theoretical nature, aims at determining adequate penalties, i.e. penalties which allow to get "oracle type inequalities" justifying the performances of the proposed procedure. A simulation study completes the theoretical results.

Mots clés: VARIABLE SELECTION; REGRESSION; BINARY CLASSIFICATION; CART

Domaines: Autre [cs.OH]

Marie Sauvé, Christine Tuleau. Variable selection through CART. [Research Report] RR-5912, INRIA. 2006. ⟨inria-00071350⟩