Inadmissibility of the corrected Akaike information criterion

DOI10.3150/23-BEJ1638MaRDI QIDQ6201858zbMATH OpenOpenAlexFDO

Authors T. Matsuda

Publication date 26 March 2024

Published in Bernoulli (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2211.09326, https://projecteuclid.org/journals/bernoulli/volume-30/issue-2/Inadmissibility-of-the-corrected-Akaike-information-criterion/10.3150/23-BEJ1638.full

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)

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.

Cites work

• scientific article; zbMATH DE number 4056770 (Why is no real title available?)
• scientific article; zbMATH DE number 18864 (Why is no real title available?)
• scientific article; zbMATH DE number 3557007 (Why is no real title available?)
• scientific article; zbMATH DE number 1964693 (Why is no real title available?)
• scientific article; zbMATH DE number 3444596 (Why is no real title available?)
• A consistency property of the AIC for multivariate linear models when the dimension and the sample size are large
• Akaike's information criterion, \(C_p\) and estimators of loss for elliptically symmetric distributions
• Ancillary Statistics and Estimation of the Loss in Estimation Problems
• Bayesian shrinkage prediction for the regression problem
• Consistency of high-dimensional AIC-type and \(C_p\)-type criteria in multivariate linear regression
• Empirical Bayes on vector observations: An extension of Stein's method
• Estimation of normal means: Frequentist estimation of loss
• Estimation under matrix quadratic loss and matrix superharmonicity
• From Fixed-X to Random-X Regression: Bias-Variance Decompositions, Covariance Penalties, and Prediction Error Estimation
• Further analysis of the data by Akaike's information criterion and the finite corrections
• Goodness of prediction fit
• IMPROVED ESTIMATION OF THE EXPECTED KULLBACK–LEIBLER DISCREPANCY IN CASE OF MISSPECIFICATION
• Improved loss estimation for a normal mean matrix
• Improved loss estimation for the lasso: a variable selection tool
• Information criteria and statistical modeling.
• Information criteria for the predictive evaluation of bayesian models
• Least Squares Model Averaging
• Least squares model averaging by Mallows criterion
• Model Selection and Multimodel Inference
• Model Selection for Multivariate Regression in Small Samples
• Modified AIC and Cp in multivariate linear regression
• Multivariate empirical Bayes and estimation of covariance matrices
• Multivariate reduced-rank regression
• On Bayes and unbiased estimators of loss
• On improved loss estimation for shrinkage estimators
• On unbiased and improved loss estimation for the mean of a multivariate normal distribution with unknown variance.
• Pitman closeness properties of Bayes shrinkage procedures in estimation and prediction
• Regression and time series model selection in small samples
• Second-order Stein: SURE for SURE and other applications in high-dimensional inference
• Shrinkage estimation
• Unifying the derivations for the Akaike and corrected Akaike information criteria.

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)