Can one estimate the conditional distribution of post-model-selection estimators?

DOI10.1214/009053606000000821MaRDI QIDQ869984zbMATH OpenOpenAlexFDO

Authors Hannes Leeb, Benedikt M. Pötscher

Publication date 12 March 2007

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0702703

zbMATH Keywords

Akaike's information criterion AICconsistencyinference after model selectionlower risk boundmodel uncertaintypre-test estimatorselection of regressorsthresholdinguniform consistency

Mathematics Subject Classification ID

Statistical decision theory (62C99) Point estimation (62F10) Asymptotic properties of parametric estimators (62F12) Linear regression; mixed models (62J05) Ridge regression; shrinkage estimators (Lasso) (62J07)

Abstract: We consider the problem of estimating the conditional distribution of a post-model-selection estimator where the conditioning is on the selected model. The notion of a post-model-selection estimator here refers to the combined procedure resulting from first selecting a model (e.g., by a model selection criterion such as AIC or by a hypothesis testing procedure) and then estimating the parameters in the selected model (e.g., by least-squares or maximum likelihood), all based on the same data set. We show that it is impossible to estimate this distribution with reasonable accuracy even asymptotically. In particular, we show that no estimator for this distribution can be uniformly consistent (not even locally). This follows as a corollary to (local) minimax lower bounds on the performance of estimators for this distribution. Similar impossibility results are also obtained for the conditional distribution of linear functions (e.g., predictors) of the post-model-selection estimator.

Recommendations

• CAN ONE ESTIMATE THE UNCONDITIONAL DISTRIBUTION OF POST-MODEL-SELECTION ESTIMATORS?
• Conditional predictive inference post model selection
• THE FINITE-SAMPLE DISTRIBUTION OF POST-MODEL-SELECTION ESTIMATORS AND UNIFORM VERSUS NONUNIFORM APPROXIMATIONS
• The distribution of a linear predictor after model selection: conditional finite-sample distributions and asymptotic approximations
• Post-model-selection method for density estimation
• The distribution of estimators after model selection:large and small sample results
• Model Selection Using Conditional Densities
• What About the Posterior Distributions When the Model is Non-dominated?

Cites work

• Asymptotic properties of maximum likelihood estimators based on conditional specification
• Conditional properties of statistical procedures
• Frequentist Model Average Estimators
• scientific article; zbMATH DE number 1220667 (Why is no real title available?)
• Incorporating lag order selection uncertainty in parameter inference for AR models
• Least squares, preliminary test and Stein‐type estimation in general vector AR(p) models
• MODEL SELECTION AND INFERENCE: FACTS AND FICTION
• Model selection by multiple test procedures
• On model uncertainty and its statistical implications. Proceedings of a workshop, held in Groningen, Netherlands, September 25-26, 1986
• On preliminary test and shrinkage M-estimation in linear models
• On the harm that ignoring pretesting can cause
• PERFORMANCE LIMITS FOR ESTIMATORS OF THE RISK OR DISTRIBUTION OF SHRINKAGE-TYPE ESTIMATORS, AND SOME GENERAL LOWER RISK-BOUND RESULTS
• The distribution of a linear predictor after model selection: conditional finite-sample distributions and asymptotic approximations
• The distribution of a linear predictor after model selection: unconditional finite-sample distributions and asymptotic approximations
• The distribution of estimators after model selection:large and small sample results
• THE FINITE-SAMPLE DISTRIBUTION OF POST-MODEL-SELECTION ESTIMATORS AND UNIFORM VERSUS NONUNIFORM APPROXIMATIONS
• Variance Estimation in a Model With Gaussian Submodels

Cited in

(69)

• On the distribution of the adaptive LASSO estimator
• Confidence intervals for high-dimensional inverse covariance estimation
• Efficient and adaptive post-model-selection estimators
• Model selection and model averaging after multiple imputation
• Uniform asymptotic inference and the bootstrap after model selection
• Sparse linear models and \(l_1\)-regularized 2SLS with high-dimensional endogenous regressors and instruments
• A penalized likelihood method for structural equation modeling
• Weighted-average least squares estimation of generalized linear models
• Exact post-selection inference for the generalized Lasso path
• Jump estimation in inverse regression
• Instrument search in pseudo-likelihood approach for nonignorable nonresponse
• Variable selection In regression models using global sensitivity analysis
• Bootstrapping multiple linear regression after variable selection
• In defense of the indefensible: a very naïve approach to high-dimensional inference
• Conditional selective inference for robust regression and outlier detection using piecewise-linear homotopy continuation
• Post-model-selection inference in linear regression models: an integrated review
• Meta-analytic Gaussian network aggregation
• Frequentist model averaging under inequality constraints
• A knockoff filter for high-dimensional selective inference
• Distribution theory of the least squares averaging estimator
• Nonparametric estimation and inference under shape restrictions
• Focused information criterion and model averaging for generalized additive partial linear models
• Frequentist model averaging with missing observations
• On various confidence intervals post-model-selection
• Estimation of panel group structure models with structural breaks in group memberships and coefficients
• Exact post-selection inference, with application to the Lasso
• The benefit of group sparsity in group inference with de-biased scaled group Lasso
• Valid post-selection inference
• CAN ONE ESTIMATE THE UNCONDITIONAL DISTRIBUTION OF POST-MODEL-SELECTION ESTIMATORS?
• CORRIGENDUM: Correction to “Performance Limits for Estimators of the Risk or Distribution of Shrinkage-Type Estimators, and Some General Lower Risk-Bound Results”
• Model averaging procedure for partially linear single-index models
• The distribution of estimators after model selection:large and small sample results
• Concentration inequalities and confidence bands for needlet density estimators on compact homogeneous manifolds
• THE FINITE-SAMPLE DISTRIBUTION OF POST-MODEL-SELECTION ESTIMATORS AND UNIFORM VERSUS NONUNIFORM APPROXIMATIONS
• Asymptotic post-selection inference for the Akaike information criterion
• A bootstrap recipe for post-model-selection inference under linear regression models
• Inference after model averaging in linear regression models
• On the length of post-model-selection confidence intervals conditional on polyhedral constraints
• Finite-sample results for lasso and stepwise Neyman-orthogonal Poisson estimators
• Valid post-selection inference in high-dimensional approximately sparse quantile regression models
• What About the Posterior Distributions When the Model is Non-dominated?
• MODEL SELECTION AND INFERENCE: FACTS AND FICTION
• Confidence intervals for low dimensional parameters in high dimensional linear models
• Weighted-average least squares prediction
• Small area estimation of the homeless in Los Angeles: an application of cost-sensitive stochastic gradient boosting
• UNIFORM-IN-SUBMODEL BOUNDS FOR LINEAR REGRESSION IN A MODEL-FREE FRAMEWORK
• Forward-selected panel data approach for program evaluation
• Approximate Selective Inference via Maximum Likelihood
• Parametric or nonparametric? A parametricness index for model selection
• A (tight) upper bound for the length of confidence intervals with conditional coverage
• Ensuring valid inference for Cox hazard ratios after variable selection
• Jackknife model averaging for composite quantile regression
• On adaptive inference and confidence bands
• Principled selection of baseline covariates to account for censoring in randomized trials with a survival endpoint
• Bounded p values in parametric programming-based selective inference
• The ET interview: Benedikt M. Pötscher
• Causal machine learning methods and use of cross-fitting in settings with high-dimensional confounding
• Causal mediation analysis: selection with asymptotically valid inference
• High dimensional binary choice model with unknown heteroskedasticity or instrumental variables
• A first-stage representation for instrumental variables quantile regression
• Post-selection inference for the Cox model with interval-censored data
• Model selection and inference for estimation of causal parameters
• Two-stage non Gaussian QML estimation of GARCH models and testing the efficiency of the Gaussian QMLE
• On the uniform asymptotic validity of subsampling and the bootstrap
• Conditional predictive inference post model selection
• On the distribution of penalized maximum likelihood estimators: the LASSO, SCAD, and thresholding
• Weighted average least squares estimation with nonspherical disturbances and an application to the Hong Kong housing market
• Approximating data
• Model averaging for semiparametric additive partial linear models

This page was built for publication: Can one estimate the conditional distribution of post-model-selection estimators?

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q869984)