Geometry of log-concave density estimation

DOI10.1007/S00454-018-0024-YMaRDI QIDQ1739202zbMATH OpenOpenAlexFDO

Authors Elina Robeva, Bernd Sturmfels, Caroline Uhler

Publication date 25 April 2019

Published in Discrete \& Computational Geometry (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1704.01910

zbMATH Keywords

non-parametric density estimationlog-concave density estimationsecondary polytopepolyhedral subdivision

Mathematics Subject Classification ID

Density estimation (62G07) Asymptotic properties of nonparametric inference (62G20) Estimation in multivariate analysis (62H12) Special polytopes (linear programming, centrally symmetric, etc.) (52B12)

Abstract: Shape-constrained density estimation is an important topic in mathematical statistics. We focus on densities on $mathbb{R}^d$ that are log-concave, and we study geometric properties of the maximum likelihood estimator (MLE) for weighted samples. Cule, Samworth, and Stewart showed that the logarithm of the optimal log-concave density is piecewise linear and supported on a regular subdivision of the samples. This defines a map from the space of weights to the set of regular subdivisions of the samples, i.e. the face poset of their secondary polytope. We prove that this map is surjective. In fact, every regular subdivision arises in the MLE for some set of weights with positive probability, but coarser subdivisions appear to be more likely to arise than finer ones. To quantify these results, we introduce a continuous version of the secondary polytope, whose dual we name the Samworth body. This article establishes a new link between geometric combinatorics and nonparametric statistics, and it suggests numerous open problems.

Recommendations

• Theoretical properties of the log-concave maximum likelihood estimator of a multidimensional density
• Maximum likelihood estimation of a multi-dimensional log-concave density. With discussion and authors' reply
• Adaptation in multivariate log-concave density estimation
• A computational approach to log-concave density estimation
• Quasi-concave density estimation

Cites work

• scientific article; zbMATH DE number 3126936 (Why is no real title available?)
• scientific article; zbMATH DE number 575960 (Why is no real title available?)
• A geometric lower bound theorem
• Assessing log-concavity of multivariate densities
• Computing the volume, counting integral points, and exponential sums
• Estimation of a convex function: Characterizations and asymptotic theory.
• Inference and modeling with log-concave distributions
• Lectures in geometric combinatorics
• Logconcavity versus logconvexity: A complete characterization
• Maximum likelihood estimation of a log-concave density and its distribution function: basic properties and uniform consistency
• Smoothed log-concave maximum likelihood estimation with applications
• The polytope of all triangulations of a point configuration
• Theoretical properties of the log-concave maximum likelihood estimator of a multidimensional density
• Total positivity, spherical series, and hypergeometric functions of matrix argument
• Triangulations. Structures for algorithms and applications

Cited in

(8)

• Total positivity in exponential families with application to binary variables
• Exact solutions in log-concave maximum likelihood estimation
• Global convergence of the log-concave MLE when the true distribution is geometric
• Statistical inference for geometric processes with lognormal distribution.
• scientific article; zbMATH DE number 5283659 (Why is no real title available?)
• Corrigendum to: ``Exact solutions in log-concave maximum likelihood estimation
• Geodesic Convexity and Covariance Estimation
• Optimal rates for estimation of two-dimensional totally positive distributions

Describes a project that uses

Uses Software

• pyuvdata
• LogConcDEAD

This page was built for publication: Geometry of log-concave density estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1739202)