Supermix: sparse regularization for mixtures

DOI10.1214/20-AOS2022zbMATH Open1475.62129arXiv1907.10592OpenAlexW3191790576MaRDI QIDQ820833FDOQ820833

Authors: Yohann De Castro, Sébastien Gadat, Clément Marteau, Cathy Maugis-Rabusseau 

Publication date: 28 September 2021

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

Abstract: This paper investigates the statistical estimation of a discrete mixing measure $mu$0 involved in a kernel mixture model. Using some recent advances in l1-regularization over the space of measures, we introduce a "data fitting and regularization" convex program for estimating $mu$0 in a grid-less manner from a sample of mixture law, this method is referred to as Beurling-LASSO. Our contribution is twofold: we derive a lower bound on the bandwidth of our data fitting term depending only on the support of $mu$0 and its so-called "minimum separation" to ensure quantitative support localization error bounds; and under a so-called "non-degenerate source condition" we derive a non-asymptotic support stability property. This latter shows that for a sufficiently large sample size n, our estimator has exactly as many weighted Dirac masses as the target $mu$0 , converging in amplitude and localization towards the true ones. Finally, we also introduce some tractable algorithms for solving this convex program based on "Sliding Frank-Wolfe" or "Conic Particle Gradient Descent". Statistical performances of this estimator are investigated designing a so-called "dual certificate", which is appropriate to our setting. Some classical situations, as e.g. mixtures of super-smooth distributions (e.g. Gaussian distributions) or ordinary-smooth distributions (e.g. Laplace distributions), are discussed at the end of the paper.

Full work available at URL: https://arxiv.org/abs/1907.10592

Recommendations

• Stable separation and super-resolution of mixture models
• A deconvolution path for mixtures
• Support recovery for sparse super-resolution of positive measures
• Multi-kernel unmixing and super-resolution using the modified matrix pencil method
• The Support Reduction Algorithm for Computing Non‐Parametric Function Estimates in Mixture Models

zbMATH Keywords

dual certificatesuperresolutionkernel approachBeurling Lassomixture recovery

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Convex programming (90C25) Numerical methods involving duality (49M29)

Cites Work

• Finite mixture models
• On the convergence properties of the EM algorithm
• Finite mixture and Markov switching models.
• Efficient maximum likelihood estimation in semiparametric mixture models
• Title not available (Why is that?)
• Title not available (Why is that?)
• Universality, Characteristic Kernels and RKHS Embedding of Measures
• Title not available (Why is that?)
• On the optimal rates of convergence for nonparametric deconvolution problems
• Semiparametric estimation of a two-component mixture model
• Statistical guarantees for the EM algorithm: from population to sample-based analysis
• Convergence of latent mixing measures in finite and infinite mixture models
• A non asymptotic penalized criterion for Gaussian mixture model selection
• Deconvolution problems in nonparametric statistics
• Maximum Likelihood: An Introduction
• Towards a Mathematical Theory of Super‐resolution
• Spike detection from inaccurate samplings
• Exact reconstruction using Beurling minimal extrapolation
• Atomic Norm Denoising With Applications to Line Spectral Estimation
• Exact support recovery for sparse spikes deconvolution
• SPADES and mixture models
• Exact Solutions to Super Resolution on Semi-Algebraic Domains in Higher Dimensions
• Semiparametric Mixtures of Symmetric Distributions
• Strong identifiability and optimal minimax rates for finite mixture estimation
• Singularity, misspecification and the convergence rate of EM
• Parameter recovery in two-component contamination mixtures: the \(L^2\) strategy
• The sliding Frank-Wolfe algorithm and its application to super-resolution microscopy

Cited In (3)

• A convex program for mixed linear regression with a recovery guarantee for well-separated data
• Localization of point scatterers via sparse optimization on measures
• Simultaneous off-the-grid learning of mixtures issued from a continuous dictionary

This page was built for publication: Supermix: sparse regularization for mixtures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q820833)