The horseshoe+ estimator of ultra-sparse signals

DOI10.1214/16-BA1028zbMATH Open1384.62079arXiv1502.00560OpenAlexW2962978766WikidataQ104437576 ScholiaQ104437576MaRDI QIDQ1699709FDOQ1699709

Authors: Anindya Bhadra, Jyotishka Datta, Nicholas G. Polson, Brandon T. Willard 

Publication date: 23 February 2018

Published in: Bayesian Analysis (Search for Journal in Brave)

Abstract: We propose a new prior for ultra-sparse signal detection that we term the "horseshoe+ prior." The horseshoe+ prior is a natural extension of the horseshoe prior that has achieved success in the estimation and detection of sparse signals and has been shown to possess a number of desirable theoretical properties while enjoying computational feasibility in high dimensions. The horseshoe+ prior builds upon these advantages. Our work proves that the horseshoe+ posterior concentrates at a rate faster than that of the horseshoe in the Kullback-Leibler (K-L) sense. We also establish theoretically that the proposed estimator has lower posterior mean squared error in estimating signals compared to the horseshoe and achieves the optimal Bayes risk in testing up to a constant. For global-local scale mixture priors, we develop a new technique for analyzing the marginal sparse prior densities using the class of Meijer-G functions. In simulations, the horseshoe+ estimator demonstrates superior performance in a standard design setting against competing methods, including the horseshoe and Dirichlet-Laplace estimators. We conclude with an illustration on a prostate cancer data set and by pointing out some directions for future research.

Full work available at URL: https://arxiv.org/abs/1502.00560

Recommendations

• The horseshoe estimator for sparse signals
• The horseshoe estimator: posterior concentration around nearly black vectors
• Adaptive posterior contraction rates for the horseshoe
• Sparsity information and regularization in the horseshoe and other shrinkage priors
• Uncertainty quantification for the horseshoe (with discussion)

zbMATH Keywords

horseshoeBayesiansparsityhorseshoe+normal meansglobal-local shrinkage

Mathematics Subject Classification ID

Asymptotic properties of parametric estimators (62F12) Bayesian inference (62F15) Bayesian problems; characterization of Bayes procedures (62C10)

Cited In (47)

• Joint mean-covariance estimation via the horseshoe
• Horseshoe Regularisation for Machine Learning in Complex and Deep Models¹
• Variance prior forms for high-dimensional Bayesian variable selection
• The horseshoe prior for time-varying parameter VARs and monetary policy
• The horseshoe estimator: posterior concentration around nearly black vectors
• Shrinkage priors for Bayesian penalized regression
• Bayesian approaches to variable selection: a comparative study from practical perspectives
• Conditions for posterior contraction in the sparse normal means problem
• Projective inference in high-dimensional problems: prediction and feature selection
• Bayesian \(l_0\)-regularized least squares
• Bayesian estimation of subset threshold autoregressions: short-term forecasting of traffic occupancy
• A flexible Bayesian variable selection approach for modeling interval data
• Contraction properties of shrinkage priors in logistic regression
• Uncertainty quantification for the horseshoe (with discussion)
• Empirical Bayes inference in sparse high-dimensional generalized linear models
• Locally adaptive smoothing with Markov random fields and shrinkage priors
• Bayesian sparse convex clustering via global-local shrinkage priors
• Bayesian Regression Using a Prior on the Model Fit: The R2-D2 Shrinkage Prior
• Bayesian regularization: from Tikhonov to horseshoe
• The horseshoe estimator for sparse signals
• Large-scale multiple hypothesis testing with the normal-beta prime prior
• Scalable multiple network inference with the joint graphical horseshoe
• Contraction of a quasi-Bayesian model with shrinkage priors in precision matrix estimation
• Bayesian shrinkage towards sharp minimaxity
• An exact sampler for fully Baysian elastic net
• Using History Matching for Prior Choice
• The horseshoe-like regularization for feature subset selection
• Bayesian discriminant analysis using a high dimensional predictor
• Sparse linear mixed model selection via streamlined variational Bayes
• Horseshoe shrinkage methods for Bayesian fusion estimation
• Bayesian sparse seemingly unrelated regressions model with variable selection and covariance estimation via the horseshoe+
• Lasso meets horseshoe: a survey
• The adaptive normal-hypergeometric-inverted-beta priors for sparse signals
• Title not available (Why is that?)
• A Scalable Empirical Bayes Approach to Variable Selection in Generalized Linear Models
• Ultra high-dimensional multivariate posterior contraction rate under shrinkage priors
• Shrinkage estimation of fixed and random effects in linear quantile mixed models
• Sparsity information and regularization in the horseshoe and other shrinkage priors
• Global-local mixtures: a unifying framework
• A novel Bayesian approach for variable selection in linear regression models
• Variable selection via penalized credible regions with Dirichlet-Laplace global-local shrinkage priors
• High-dimensional multivariate posterior consistency under global-local shrinkage priors
• Shared Bayesian variable shrinkage in multinomial logistic regression
• A Bayesian model of dose-response for cancer drug studies
• The Graphical Horseshoe Estimator for Inverse Covariance Matrices
• Data-driven priors and their posterior concentration rates
• Prediction risk for the horseshoe regression

Uses Software

• Stan

This page was built for publication: The horseshoe+ estimator of ultra-sparse signals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1699709)