Weak convergence of the regularization path in penalized M-estimation

DOI10.1111/J.1467-9469.2009.00682.XzbMATH Open1226.60032arXiv0809.0064OpenAlexW1592037586MaRDI QIDQ3103136FDOQ3103136

François Roueff, Jean-Francois Germain

Publication date: 26 November 2011

Published in: Scandinavian Journal of Statistics (Search for Journal in Brave)

Abstract: We consider an estimator $hbbet{a}_n(t)$ defined as the element $$ minimizing a contrast process $$ for each t. We give some general results for deriving the weak convergence of $$ in the space of bounded functions, where, for each t, $$ is the $$ minimizing the limit of $$ as $noinfty$. These results are applied in the context of penalized M-estimation, that is, when $$, where $M_n$ is a usual contrast process and $J_n$ a penalty such as the $el{l}^1$ norm or the squared $el{l}^2$ norm. The function $hbbet{a}_n$ is then called a emph{regularization path}. For instance we show that the central limit theorem established for the lasso estimator in Knight and Fu (2000) continues to hold in a functional sense for the regularization path. Other examples include various possible contrast processes for $M_n$ such as those considered in Pollard (1985).

Full work available at URL: https://arxiv.org/abs/0809.0064

Recommendations

• A generic path algorithm for regularized statistical estimation
• On the structure of regularization paths for piecewise differentiable regularization terms
• Piecewise linear regularized solution paths
• The Penalized Analytic Center Estimator
• Weakly decomposable regularization penalties and structured sparsity

zbMATH Keywords

lassoweak convergenceregularization pathAkaike information criterion (AIC)pathwise argmin theorempenalized M-estimation

Mathematics Subject Classification ID

Asymptotic properties of nonparametric inference (62G20) Ridge regression; shrinkage estimators (Lasso) (62J07) Central limit and other weak theorems (60F05) Functional limit theorems; invariance principles (60F17)

Cites Work

• Least angle regression. (With discussion)
• Weak convergence and empirical processes. With applications to statistics
• Asymptotic Statistics
• Title not available (Why is that?)
• Title not available (Why is that?)
• Simultaneous analysis of Lasso and Dantzig selector
• Title not available (Why is that?)
• Ridge Regression: Biased Estimation for Nonorthogonal Problems
• Asymptotics for Lasso-type estimators.
• Sparsity oracle inequalities for the Lasso
• On the ``degrees of freedom of the lasso
• Title not available (Why is that?)
• Asymptotics for \(M\)-estimators defined by convex minimization

Cited In (2)

• A note on the asymptotic distribution of lasso estimator for correlated data
• Sparse least trimmed squares regression for analyzing high-dimensional large data sets

This page was built for publication: Weak convergence of the regularization path in penalized M-estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3103136)