Inadmissibility of the corrected Akaike information criterion

DOI10.3150/23-BEJ1638arXiv2211.09326OpenAlexW4391458284MaRDI QIDQ6201858FDOQ6201858

Authors: T. Matsuda 

Publication date: 26 March 2024

Published in: Bernoulli (Search for Journal in Brave)

Abstract: For the multivariate linear regression model with unknown covariance, the corrected Akaike information criterion is the minimum variance unbiased estimator of the expected Kullback--Leibler discrepancy. In this study, based on the loss estimation framework, we show its inadmissibility as an estimator of the Kullback--Leibler discrepancy itself, instead of the expected Kullback--Leibler discrepancy. We provide improved estimators of the Kullback--Leibler discrepancy that work well in reduced-rank situations and examine their performance numerically.

Full work available at URL: https://arxiv.org/abs/2211.09326

zbMATH Keywords

Akaike information criterionadmissibilityloss estimationKullback-Leibler discrepancycorrected Akaike information criterion

Mathematics Subject Classification ID

Statistics (62-XX) Information and communication theory, circuits (94-XX)

Cites Work

• Regression and time series model selection in small samples
• Title not available (Why is that?)
• Title not available (Why is that?)
• Multivariate reduced-rank regression
• Model Selection and Multimodel Inference
• Further analysis of the data by Akaike's information criterion and the finite corrections
• Bayesian shrinkage prediction for the regression problem
• Information criteria and statistical modeling.
• Multivariate empirical Bayes and estimation of covariance matrices
• On Bayes and unbiased estimators of loss
• Title not available (Why is that?)
• Goodness of prediction fit
• Model Selection for Multivariate Regression in Small Samples
• Least squares model averaging by Mallows criterion
• Least Squares Model Averaging
• Empirical Bayes on vector observations: An extension of Stein's method
• Consistency of high-dimensional AIC-type and \(C_p\)-type criteria in multivariate linear regression
• Modified AIC and Cp in multivariate linear regression
• Pitman closeness properties of Bayes shrinkage procedures in estimation and prediction
• Title not available (Why is that?)
• Unifying the derivations for the Akaike and corrected Akaike information criteria.
• A consistency property of the AIC for multivariate linear models when the dimension and the sample size are large
• IMPROVED ESTIMATION OF THE EXPECTED KULLBACK–LEIBLER DISCREPANCY IN CASE OF MISSPECIFICATION
• On unbiased and improved loss estimation for the mean of a multivariate normal distribution with unknown variance.
• Title not available (Why is that?)
• Ancillary Statistics and Estimation of the Loss in Estimation Problems
• Estimation of normal means: Frequentist estimation of loss
• Estimation under matrix quadratic loss and matrix superharmonicity
• On improved loss estimation for shrinkage estimators
• Information criteria for the predictive evaluation of bayesian models
• Improved loss estimation for a normal mean matrix
• Second-order Stein: SURE for SURE and other applications in high-dimensional inference
• Improved loss estimation for the lasso: a variable selection tool
• Akaike's information criterion, \(C_p\) and estimators of loss for elliptically symmetric distributions
• Shrinkage estimation
• From Fixed-X to Random-X Regression: Bias-Variance Decompositions, Covariance Penalties, and Prediction Error Estimation

This page was built for publication: Inadmissibility of the corrected Akaike information criterion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6201858)