Inference for Local Parameters in Convexity Constrained Models

DOI10.1080/01621459.2022.2071721MaRDI QIDQ6185569zbMATH OpenOpenAlexFDO

Authors Hang Deng, Qiyang Han, Bodhisattva Sen

Publication date 8 January 2024

Published in Journal of the American Statistical Association (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2006.10264, https://figshare.com/articles/journal_contribution/Inference_for_local_parameters_in_convexity_constrained_models/19686439

zbMATH Keywords

convex regressionshape constraintslog-concave density estimationconvex least squares estimatorpivotal limit distribution theory

Mathematics Subject Classification ID

Statistics (62-XX)

Abstract: We consider the problem of inference for local parameters of a convex regression function $f_0:[0,1]omathbbR$ based on observations from a standard nonparametric regression model, using the convex least squares estimator (LSE) $widehat{f}_n$. For $x_{0}in(0,1)$, the local parameters include the pointwise function value $f_0(x_0)$, the pointwise derivative ${f_0}^{\prime }(x_0)$, and the anti-mode (i.e., the smallest minimizer) of $f_0$. The existing limiting distribution of the estimation error $(widehat{f}_n(x_0)-f_0(x_0),widehat{f^{\prime }}_n(x_0)-{f_0}^{\prime }(x_0))$ depends on the unknown second derivative ${f_0}^{″}(x_0)$, and is therefore not directly applicable for inference. To circumvent this impasse, we show that the following locally normalized errors (LNEs) enjoy pivotal limiting behavior: Let $[widehatu(x_0),widehatv(x_0)]$ be the maximal interval containing $x_0$ where $widehat{f}_n$ is linear. Then, under standard conditions, ${\displaystyle }$�inom{ sqrt{n(widehat{v}(x_0)-widehat{u}(x_0))}(widehat{f}_n(x_0)-f_0(x_0)) }{ sqrt{n(widehat{v}(x_0)-widehat{u}(x_0))^3}(widehat{f}_n'(x_0)-f_0'(x_0))} ightsquigarrow sigma cdot �inom{mathbb{L}^{(0)}_2}{mathbb{L}^{(1)}_2},${\displaystyle }$ where $n$ is the sample size, $sigma$ is the standard deviation of the errors, and $mathbb{L}_2^{(0)},mathbb{L}_2^{(1)}$ are universal random variables. This asymptotically pivotal LNE theory instantly yields a simple tuning-free procedure for constructing CIs with asymptotically exact coverage and optimal length for $f_0(x_0)$ and ${f_0}^{\prime }(x_0)$. We also construct an asymptotically pivotal LNE for the anti-mode of $f_0$, and its limiting distribution does not even depend on $sigma$. These asymptotically pivotal LNE theories are further extended to other convexity/concavity constrained models (e.g., log-concave density estimation) for which a limit distribution theory is available for problem-specific estimators.

Cites work

• scientific article; zbMATH DE number 193483 (Why is no real title available?)
• scientific article; zbMATH DE number 1862807 (Why is no real title available?)
• A canonical process for estimation of convex functions: the ``invelope of integrated Brownian motion \(+t^ 4\).
• Adaptation in log-concave density estimation
• Adaptation in multivariate log-concave density estimation
• Adaptive confidence intervals for regression functions under shape constraints
• An Effective Bandwidth Selector for Local Least Squares Regression
• Approximation and estimation of s-concave densities via Rényi divergences
• Approximation by log-concave distributions, with applications to regression
• Bandwidth choice for nonparametric regression
• Change-point in stochastic design regression and the bootstrap
• Confidence intervals for multiple isotonic regression and other monotone models
• Consistency in concave regression
• Consistency of concave regression with an application to current-status data
• Detecting the Presence of Mixing with Multiscale Maximum Likelihood
• Distribution of the Ratio of the Mean Square Successive Difference to the Variance
• Effect of mean on variance function estimation in nonparametric regression
• Estimating a concave distribution function from data corrupted with additive noise
• Estimation of a convex function: Characterizations and asymptotic theory.
• Global rates of convergence in log-concave density estimation
• Global rates of convergence of the MLEs of log-concave and \(s\)-concave densities
• Global risk bounds and adaptation in univariate convex regression
• Inconsistency of bootstrap: the Grenander estimator
• Independent component analysis via nonparametric maximum likelihood estimation
• Inference for a two-component mixture of symmetric distributions under log-concavity
• Inference for the mode of a log-concave density
• Limit distribution theory for maximum likelihood estimation of a log-concave density
• Local continuity of log-concave projection, with applications to estimation under model misspecification
• Log-concavity and strong log-concavity: a review
• Maximum likelihood estimation of a log-concave density and its distribution function: basic properties and uniform consistency
• Maximum likelihood estimation of a multi-dimensional log-concave density. With discussion and authors' reply
• Nonparametric estimation of a convex bathtub-shaped hazard function
• Nonparametric estimation of multivariate convex-transformed densities
• Nonparametric option pricing under shape restrictions
• Nonparametric regression under qualitative smoothness assumptions
• On Difference-Based Variance Estimation in Nonparametric Regression When the Covariate is High Dimensional
• On risk bounds in isotonic and other shape restricted regression problems
• On univariate convex regression
• Optimal estimation of variance in nonparametric regression with random design
• Point Estimates of Ordinates of Concave Functions
• Quasi-concave density estimation
• Recent progress in log-concave density estimation
• Semiparametric Estimation of Monotone and Concave Utility Functions for Polychotomous Choice Models
• Set structured global empirical risk minimizers are rate optimal in general dimensions
• Sharp oracle inequalities for least squares estimators in shape restricted regression
• Smooth tail-index estimation
• Smoothed log-concave maximum likelihood estimation with applications
• The Nonparametric Approach to Production Analysis
• Theoretical properties of the log-concave maximum likelihood estimator of a multidimensional density

Cited in

(2)

• Local continuity of log-concave projection, with applications to estimation under model misspecification
• Estimation and inference for minimizer and minimum of convex functions: optimality, adaptivity and uncertainty principles

This page was built for publication: Inference for Local Parameters in Convexity Constrained Models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6185569)