Sharp convergence rate of the estimation error of LASSO via the Feature Space Decomposition method
In the authors' words
In this paper, we use the Feature Space Decomposition (FSD) method to establish an explicit, non-asymptotic, instance-optimal convergence rate for the estimation error of the LASSO in the Gaussian linear regression problem for any signal, any diagonal covariance and any regularization parameter above where is the noise variance and the number of data.
Main resultThe abstract does not state a limitation.
Appeared: Friday, October 23. arXiv. Preprint, not yet peer-reviewed.