Adaptation to lowest density regions with application to support recovery

DOI10.1214/15-AOS1366zbMATH Open1331.62222arXiv1408.4057MaRDI QIDQ5963522FDOQ5963522

Angelika Rohde, Tim Patschkowski

Publication date: 22 February 2016

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

Abstract: A scheme for locally adaptive bandwidth selection is proposed which sensitively shrinks the bandwidth of a kernel estimator at lowest density regions such as the support boundary which are unknown to the statistician. In case of a H"{o}lder continuous density, this locally minimax-optimal bandwidth is shown to be smaller than the usual rate, even in case of homogeneous smoothness. Some new type of risk bound with respect to a density-dependent standardized loss of this estimator is established. This bound is fully nonasymptotic and allows to deduce convergence rates at lowest density regions that can be substantially faster than $n^{-1/2}$. It is complemented by a weighted minimax lower bound which splits into two regimes depending on the value of the density. The new estimator adapts into the second regime, and it is shown that simultaneous adaptation into the fastest regime is not possible in principle as long as the H"{o}lder exponent is unknown. Consequences on plug-in rules for support recovery are worked out in detail. In contrast to those with classical density estimators, the plug-in rules based on the new construction are minimax-optimal, up to some logarithmic factor.

Full work available at URL: https://arxiv.org/abs/1408.4057

Recommendations

• New approaches to locally adaptive nonparametric estimation and inference
• A plug-in approach to support estimation
• Adaptive density estimation: A curse of support?
• Lower bounds for bandwidth selection in density estimation
• Bandwidth selection in kernel density estimation: oracle inequalities and adaptive minimax optimality

zbMATH Keywords

bandwidth selectionsupport estimationanisotropic density estimationadaptation to lowest density regionsdensity dependent minimax optimality

Mathematics Subject Classification ID

Density estimation (62G07)

Cites Work

• Confidence bands in density estimation
• A plug-in approach to support estimation
• Asymptotical minimax recovery of sets with smooth boundaries
• Detection of Abnormal Behavior Via Nonparametric Estimation of the Support
• On boundary estimation
• On nonparametric estimation of density level sets
• Title not available (Why is that?)
• Title not available (Why is that?)
• Optimal rates for plug-in estimators of density level sets
• Smooth discrimination analysis
• Measuring mass concentrations and estimating density contour clusters -- An excess mass approach
• A robust, adaptive M-estimator for pointwise estimation in heteroscedastic regression
• Title not available (Why is that?)
• Asymptotic equivalence of density estimation and Gaussian white noise
• Estimation of functionals of density support
• On minimax density estimation on \(\mathbb R\)
• Isoperimetric inequalities. Differential geometric and analytic perspectives
• On adaptive minimax density estimation on \(\mathbb R^d\)
• Bandwidth selection in kernel density estimation: oracle inequalities and adaptive minimax optimality
• Adaptive density estimation: A curse of support?
• Estimation of non-sharp support boundaries
• On Poincaré cone property
• Set estimation and nonparametric detection.
• Title not available (Why is that?)
• Nonlinear estimation in anisotropic multi-index denoising
• Minimax theory of image reconstruction
• Asymptotic normality in density support estimation
• Optimal aggregation of classifiers in statistical learning.
• Model selection for density estimation with \(\mathbb L_2\)-loss
• Adaptive pointwise estimation of conditional density function
• Trade-offs between global and local risks in nonparametric function estimation
• �ber die konvexe H�lle von n zuf�llig gew�hlten Punkten. II
• �ber die konvexe H�lle von n zuf�llig gew�hlten Punkten
• Multivariate density estimation under sup-norm loss: oracle approach, adaptation and independence structure
• Spatially adaptive density estimation by localised Haar projections
• Adaptive estimation of and oracle inequalities for probability densities and characteristic functions
• Title not available (Why is that?)
• Exact rates in density support estimation
• Minimax and minimax adaptive estimation in multiplicative regression: locally Bayesian approach
• Exact adaptive pointwise estimation on Sobolev classes of densities
• On estimating the endpoint of a distribution
• Calibration of self-decomposable Lévy models
• Adaptive function estimation in nonparametric regression with one-sided errors
• Limit theorems for convex hulls
• On the estimation of a support curve of indeterminate sharpness
• Optimal calibration for multiple testing against local inhomogeneity in higher dimension
• Adaptive goodness-of-fit tests based on signed ranks
• Adaptive estimation of convex polytopes and convex sets from noisy data
• Adaptive estimation over anisotropic functional classes via oracle approach
• Complexity penalized support estimation.
• Adaptation to lowest density regions with application to support recovery
• Optimal estimation accuracy of nonsmooth images
• Title not available (Why is that?)

Cited In (11)

• A regularity class for the roots of nonnegative functions
• Local convergence rates of the nonparametric least squares estimator with applications to transfer learning
• Local bandwidth selection for kernel density estimation in a bifurcating Markov chain model
• Optimal rates of entropy estimation over Lipschitz balls
• The empirical Christoffel function with applications in data analysis
• The Le Cam distance between density estimation, Poisson processes and Gaussian white noise
• Statistics with set-valued functions: applications to inverse approximate optimization
• Asymptotic nonequivalence of density estimation and Gaussian white noise for small densities
• Adaptation to lowest density regions with application to support recovery
• Optimization hierarchy for fair statistical decision problems
• Rate of convergence for geometric inference based on the empirical Christoffel function

This page was built for publication: Adaptation to lowest density regions with application to support recovery

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5963522)