Asymptotic minimaxity of false discovery rate thresholding for sparse exponential data

DOI10.1214/009053606000000920zbMATH Open1114.62010arXivmath/0602311OpenAlexW3105074488WikidataQ29308446 ScholiaQ29308446MaRDI QIDQ2373587FDOQ2373587

Authors: David Donoho, Jiashun Jin 

Publication date: 12 July 2007

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

Abstract: We apply FDR thresholding to a non-Gaussian vector whose coordinates X_i, i=1,..., n, are independent exponential with individual means $m{u}_i$. The vector $mu=(m{u}_i)$ is thought to be sparse, with most coordinates 1 but a small fraction significantly larger than 1; roughly, most coordinates are simply `noise,' but a small fraction contain `signal.' We measure risk by per-coordinate mean-squared error in recovering $log(m{u}_i)$, and study minimax estimation over parameter spaces defined by constraints on the per-coordinate p-norm of $log(m{u}_i)$: $frac1nsu{m}_{i=1}^{n}lo{g}^p(m{u}_i)leqet{a}^p$. We show for large n and small $eta$ that FDR thresholding can be nearly Minimax. The FDR control parameter 0<q<1 plays an important role: when $qleq1/2$, the FDR estimator is nearly minimax, while choosing a fixed q>1/2 prevents near minimaxity. These conclusions mirror those found in the Gaussian case in Abramovich et al. [Ann. Statist. 34 (2006) 584--653]. The techniques developed here seem applicable to a wide range of other distributional assumptions, other loss measures and non-i.i.d. dependency structures.

Full work available at URL: https://arxiv.org/abs/math/0602311

Recommendations

• Adapting to unknown sparsity by controlling the false discovery rate
• On false discovery rate thresholding for classification under sparsity
• Some optimality properties of FDR controlling rules under sparsity
• Asymptotic Bayes-optimality under sparsity of some multiple testing procedures
• Strong Control, Conservative Point Estimation and Simultaneous Conservative Consistency of False Discovery Rates: A Unified Approach

zbMATH Keywords

multiple comparisonssparsitythreshold rulesminimum Bayes estimationmixtures of exponential models

Mathematics Subject Classification ID

Asymptotic properties of nonparametric inference (62G20) Estimation in multivariate analysis (62H12) Bayesian problems; characterization of Bayes procedures (62C10) Empirical decision procedures; empirical Bayes procedures (62C12) Minimax procedures in statistical decision theory (62C20) Paired and multiple comparisons; multiple testing (62J15)

Cites Work

• Title not available (Why is that?)
• An improved Bonferroni procedure for multiple tests of significance
• Asymptotic minimaxity of false discovery rate thresholding for sparse exponential data
• Adapting to unknown sparsity by controlling the false discovery rate
• The tight constant in the Dvoretzky-Kiefer-Wolfowitz inequality
• Minimax risk over \(l_ p\)-balls for \(l_ q\)-error
• Title not available (Why is that?)
• Asymptotic Minimax Character of the Sample Distribution Function and of the Classical Multinomial Estimator
• Title not available (Why is that?)
• Minimax estimation via wavelet shrinkage
• Operating Characteristics and Extensions of the False Discovery Rate Procedure
• The Power of Rank Tests
• Title not available (Why is that?)
• Title not available (Why is that?)

Cited In (16)

• False discovery rate control with unknown null distribution: is it possible to mimic the oracle?
• Asymptotic Bayes-optimality under sparsity of some multiple testing procedures
• Optimal rates of convergence for estimating the null density and proportion of nonnull effects in large-scale multiple testing
• Asymptotic minimaxity of false discovery rate thresholding for sparse exponential data
• Properties of higher criticism under strong dependence
• Adapting to unknown sparsity by controlling the false discovery rate
• Feature selection by higher criticism thresholding achieves the optimal phase diagram
• Proportion of Non-Zero Normal Means: Universal Oracle Equivalences and Uniformly Consistent Estimators
• On false discovery rate thresholding for classification under sparsity
• Rate of convergence of a risk estimator to the normal law in a multiple hypothesis testing problem using the FDR threshold
• Innovated higher criticism for detecting sparse signals in correlated noise
• Discovering the False Discovery Rate
• Asymptotically optimal sequential FDR and pFDR control with (or without) prior information on the number of signals
• Some optimality properties of FDR controlling rules under sparsity
• Modified versions of the Bayesian information criterion for sparse generalized linear models
• On the false discovery rate and an asymptotically optimal rejection curve

This page was built for publication: Asymptotic minimaxity of false discovery rate thresholding for sparse exponential data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2373587)