Sharp Convergence Rates for Forward Regression in High-Dimensional\n Sparse Linear Models

2017/02/03 by Damian Kozbur, Kozbur, Damian
Engineering · Mathematics · #Control Systems and Identification #FOS: Computer and information sciences #Machine Learning (stat.ML) #Sparse and Compressive Sensing Techniques #Statistical Methods and Inference

paper · pdf · doi:10.48550/arxiv.1702.01000

openalex publication_date 2017/02/03 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28

Abstract

Forward regression is a statistical model selection and estimation procedure\nwhich inductively selects covariates that add predictive power into a working\nstatistical regression model. Once a model is selected, unknown regression\nparameters are estimated by least squares. This paper analyzes forward\nregression in high-dimensional sparse linear models. Probabilistic bounds for\nprediction error norm and number of selected covariates are proved. The\nanalysis in this paper gives sharp rates and does not require beta-min or\nirrepresentability conditions.\n

Related

• Analysis of Testing-Based Forward Model Selection
• Minimax Optimal Rates of Estimation in High Dimensional Additive Models: Universal Phase Transition
• Adaptive posterior concentration rates for sparse high-dimensional linear regression with random design and unknown error variance
• Optimal prediction for sparse linear models? Lower bounds for coordinate-separable M-estimators
• Sparse Linear Models and Two-Stage Estimation in High-Dimensional Settings with Possibly Many Endogenous Regressors