Penalized maximum likelihood estimation and effective dimension

DOI10.1214/15-AIHP720zbMATH Open1421.62021arXiv1205.0498OpenAlexW2963550177MaRDI QIDQ520783FDOQ520783

Authors: V. G. Spokoiny 

Publication date: 6 April 2017

Published in: Annales de l'Institut Henri Poincaré. Probabilités et Statistiques (Search for Journal in Brave)

Abstract: This paper extends some prominent statistical results including emph{Fisher Theorem and Wilks phenomenon} to the penalized maximum likelihood estimation with a quadratic penalization. It appears that sharp expansions for the penalized MLE ( ilde{ hetav}_{G} ) and for the penalized maximum likelihood can be obtained without involving any asymptotic arguments, the results only rely on smoothness and regularity properties of the of the considered log-likelihood function. The error of estimation is specified in terms of the effective dimension (p_G ) of the parameter set which can be much smaller than the true parameter dimension and even allows an infinite dimensional functional parameter. In the i.i.d. case, the Fisher expansion for the penalized MLE can be established under the constraint "(p_G^{2}/n) is small" while the remainder in the Wilks result is of order (p_G^{3}/n ).

Full work available at URL: https://arxiv.org/abs/1205.0498

Recommendations

• Asymptotic analysis of penalized likelihood and related estimators
• Model selection and sharp asymptotic minimaxity
• scientific article
• A nonparametric Cramér-Rao inequality for estimators of statistical functionals.

zbMATH Keywords

maximum likelihood estimationpenaltyWilks and Fisher expansions

Mathematics Subject Classification ID

Point estimation (62F10) Parametric tolerance and confidence regions (62F25) Estimation in multivariate analysis (62H12) Generalized linear models (logistic models) (62J12)

Cited In (9)

• On Dantzig and Lasso estimators of the drift in a high dimensional Ornstein-Uhlenbeck model
• The likelihood ratio test in high-dimensional logistic regression is asymptotically a rescaled Chi-square
• Empirical Bayes inference in sparse high-dimensional generalized linear models
• MM for penalized estimation
• Large ball probabilities, Gaussian comparison and anti-concentration
• Gaussian approximation for penalized Wasserstein barycenters
• Title not available (Why is that?)
• On accuracy of Gaussian approximation in Bayesian semiparametric problems
• Large sample properties of the SCAD-penalized maximum likelihood estimation on high dimen\-sions

This page was built for publication: Penalized maximum likelihood estimation and effective dimension

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q520783)