Assumptionless consistency of the Lasso

Abstract

The Lasso is a popular statistical tool invented by Robert Tibshirani for linear regression when the number of covariates is greater than or comparable to the number of observations. The purpose of this note is to highlight the simple fact (noted in a number of earlier papers in various guises) that for the loss function considered in Tibshirani's original paper, the Lasso is consistent under almost no assumptions at all.

10 pages. Typos corrected

References in corpus (1)

• The Dantzig selector and sparsity oracle inequalities

Cited by in corpus (14)

• A new perspective on least squares under convex constraint
• Prediction error of cross-validated Lasso
• Counterfactual Predictions under Runtime Confounding
• Nonparametric Regression with Adaptive Truncation via a Convex Hierarchical Penalty
• On Uniform Convergence and Low-Norm Interpolation Learning
• High dimensional regression and matrix estimation without tuning parameters
• Multimodal Neuroimaging Data Integration and Pathway Analysis
• Convergence of Parameter Estimates for Regularized Mixed Linear Regression Models
• On the Reconstruction Risk of Convolutional Sparse Dictionary Learning
• On Learning Continuous Pairwise Markov Random Fields
• Maximum Regularized Likelihood Estimators: A General Prediction Theory and Applications
• From Fixed-X to Random-X Regression: Bias-Variance Decompositions, Covariance Penalties, and Prediction Error Estimation
• Generalization Bounds for High-dimensional M-estimation under Sparsity Constraint
• The Balancing Act in Causal Inference