Krylov subspace solvers for ℓ1 regularized logistic regression method - Archive ouverte HAL Access content directly
Journal Articles Communications in Statistics - Simulation and Computation Year : 2023

Krylov subspace solvers for ℓ1 regularized logistic regression method

M. El Guide
  • Function : Author
Christos Koukouvinos
  • Function : Author
  • PersonId : 880397
A. Lappa
  • Function : Author

Abstract

In this paper, we propose an approach based on Krylov subspace methods for the solution of ℓ1 regularized logistic regression problem. The main idea is to transform the constrained ℓ2-ℓ1 minimization problem obtained by applying the IRLS method to a ℓ2-ℓ2 one that allow regularization matrices in the usual 2-norm regularization term. The regularization parameter that controls the equilibrium between the minimization of the two terms of the ℓ2-ℓ2 minimization problem can be then chosen inexpensively by solving some reduced minimization problems related to generalized cross-validation (GCV) methods. These reduced problems can be obtained after a few iterations of Krylov subspace based methods. The goal of our simulation study is directed toward the variable selection and the prediction accuracy performance of the proposed method in solving a ℓ1 regularized logistic regression problem in large dimensional data with different correlation structures among predictors. Finally, real data are used to confirm the efficiency of the proposed method in terms of the computational cost.
No file

Dates and versions

hal-04413491 , version 1 (23-01-2024)

Identifiers

Cite

M. El Guide, Khalide Jbilou, Christos Koukouvinos, A. Lappa. Krylov subspace solvers for ℓ1 regularized logistic regression method. Communications in Statistics - Simulation and Computation, 2023, 52 (6), pp.2738-2751. ⟨10.1080/03610918.2021.1914093⟩. ⟨hal-04413491⟩
8 View
0 Download

Altmetric

Share

Gmail Facebook X LinkedIn More