Multivariate medians and measure-symmetrization

Richard A. Vitale

DOI: https://doi.org/10.1214/074921707000000201

2007-08-08

Abstract:We discuss two research areas dealing respectively with (1) a class of multivariate medians and (2) a symmetrization algorithm for probability measures.

Statistics Theory

What problem does this paper attempt to address?

Median Shapes

Yunfeng Hu,Matthew Hudelson,Bala Krishnamoorthy,Altansuren Tumurbaatar,Kevin R. Vixie

DOI: https://doi.org/10.48550/arXiv.1802.04968

2018-12-10

Abstract:We introduce and begin to explore the mean and median of finite sets of shapes represented as integral currents. The median can be computed efficiently in practice, and we focus most of our theoretical and computational attention on medians. We consider questions on the existence and regularity of medians. While the median might not exist in all cases, we show that a mass-regularized median is guaranteed to exist. When the input shapes are modeled by integral currents with shared boundaries in codimension $1$, we show that the median is guaranteed to exist, and is contained in the \emph{envelope} of the input currents. On the other hand, we show that medians can be \emph{wild} in this setting, and smooth inputs can generate non-smooth medians. For higher codimensions, we show that \emph{books} are minimizing for a finite set of $1$-currents in $\Bbb{R}^3$ with shared boundaries. As part of this proof, we present a new result in graph theory---that \emph{cozy} graphs are \emph{comfortable}---which should be of independent interest. Further, we show that regular points on the median have book-like tangent cones in this case. From the point of view of computation, we study the median shape in the settings of a finite simplicial complex. When the input shapes are represented by chains of the simplicial complex, we show that the problem of finding the median shape can be formulated as an integer linear program. This optimization problem can be solved as a linear program in practice, thus allowing one to compute median shapes efficiently. We provide open source code implementing our methods, which could also be used by anyone to experiment with ideas of their own. The software could be accessed at <a class="link-external link-https" href="https://github.com/tbtraltaa/medianshape" rel="external noopener nofollow">this https URL</a>.

Differential Geometry,Computational Geometry,Algebraic Topology,Optimization and Control
On the Wasserstein Median of Probability Measures

Kisung You,Dennis Shung,Mauro Giuffrè

DOI: https://doi.org/10.1080/10618600.2024.2374580

2024-09-02

Journal of Computational and Graphical Statistics

Abstract:The primary choice to summarize a finite collection of random objects is by using measures of central tendency, such as mean and median. In the field of optimal transport, the Wasserstein barycenter corresponds to the Fréchet or geometric mean of a set of probability measures, which is defined as a minimizer of the sum of squared distances to each element in a given set with respect to the Wasserstein distance of order 2. We introduce the Wasserstein median as a robust alternative to the Wasserstein barycenter. The Wasserstein median corresponds to the Fréchet median under the 2-Wasserstein metric. The existence and consistency of the Wasserstein median are first established, along with its robustness property. In addition, we present a general computational pipeline that employs any recognized algorithms for the Wasserstein barycenter in an iterative fashion and demonstrate its convergence. The utility of the Wasserstein median as a robust measure of central tendency is demonstrated using real and simulated data.

statistics & probability
Differentially private multivariate medians

Kelly Ramsay,Aukosh Jagannath,Shoja'eddin Chenouri

2024-03-27

Abstract:Statistical tools which satisfy rigorous privacy guarantees are necessary for modern data analysis. It is well-known that robustness against contamination is linked to differential privacy. Despite this fact, using multivariate medians for differentially private and robust multivariate location estimation has not been systematically studied. We develop novel finite-sample performance guarantees for differentially private multivariate depth-based medians, which are essentially sharp. Our results cover commonly used depth functions, such as the halfspace (or Tukey) depth, spatial depth, and the integrated dual depth. We show that under Cauchy marginals, the cost of heavy-tailed location estimation outweighs the cost of privacy. We demonstrate our results numerically using a Gaussian contamination model in dimensions up to d = 100, and compare them to a state-of-the-art private mean estimation algorithm. As a by-product of our investigation, we prove concentration inequalities for the output of the exponential mechanism about the maximizer of the population objective function. This bound applies to objective functions that satisfy a mild regularity condition.

Statistics Theory,Cryptography and Security,Machine Learning
Variance Inequalities for Transformed Fréchet Means in Hadamard Spaces

Christof Schötz

2024-09-04

Abstract:The Fréchet mean (or barycenter) generalizes the expectation of a random variable to metric spaces by minimizing the expected squared distance to the random variable. Similarly, the median can be generalized by its property of minimizing the expected absolute distance. We consider the class of transformed Fréchet means with nondecreasing, convex transformations that have a concave derivative. This class includes the Fréchet median, the Fréchet mean, the Huber loss-induced Fréchet mean, and other statistics related to robust statistics in metric spaces. We study variance inequalities for these transformed Fréchet means. These inequalities describe how the expected transformed distance grows when moving away from a minimizer, i.e., from a transformed Fréchet mean. Variance inequalities are useful in the theory of estimation and numerical approximation of transformed Fréchet means. Our focus is on variance inequalities in Hadamard spaces - metric spaces with globally nonpositive curvature. Notably, some results are new also for Euclidean spaces. Additionally, we are able to characterize uniqueness of transformed Fréchet means, in particular of the Fréchet median.

Probability,Statistics Theory
Sibson's $α$-Mutual Information and its Variational Representations

Amedeo Roberto Esposito,Michael Gastpar,Ibrahim Issa

2024-05-14

Abstract:Information measures can be constructed from Rényi divergences much like mutual information from Kullback-Leibler divergence. One such information measure is known as Sibson's $\alpha$-mutual information and has received renewed attention recently in several contexts: concentration of measure under dependence, statistical learning, hypothesis testing, and estimation theory. In this paper, we survey and extend the state of the art. In particular, we introduce variational representations for Sibson's $\alpha$-mutual information and employ them in each of the contexts just described to derive novel results. Namely, we produce generalized Transportation-Cost inequalities and Fano-type inequalities. We also present an overview of known applications, spanning from learning theory and Bayesian risk to universal prediction.

Information Theory,Probability
Bayesian Inference on Multivariate Medians and Quantiles

Indrabati Bhattacharya,Subhashis Ghosal

DOI: https://doi.org/10.5705/ss.202020.0108

2019-09-23

Abstract:In this paper, we consider Bayesian inference on a class of multivariate median and the multivariate quantile functionals of a joint distribution using a Dirichlet process prior. Since, unlike univariate quantiles, the exact posterior distribution of multivariate median and multivariate quantiles are not obtainable explicitly, we study these distributions asymptotically. We derive a Bernstein-von Mises theorem for the multivariate $\ell_1$-median with respect to general $\ell_p$-norm, which in particular shows that its posterior concentrates around its true value at $n^{-1/2}$-rate and its credible sets have asymptotically correct frequentist coverage. In particular, asymptotic normality results for the empirical multivariate median with general $\ell_p$-norm is also derived in the course of the proof which extends the results from the case $p=2$ in the literature to a general $p$. The technique involves approximating the posterior Dirichlet process by a Bayesian bootstrap process and deriving a conditional Donsker theorem. We also obtain analogous results for an affine equivariant version of the multivariate $\ell_1$-median based on an adaptive transformation and re-transformation technique. The results are extended to a joint distribution of multivariate quantiles. The accuracy of the asymptotic result is confirmed by a simulation study. We also use the results to obtain Bayesian credible regions for multivariate medians for Fisher's iris data, which consists of four features measured for each of three plant species.

Statistics Theory
Generalizing the Mean and Variance to Categorical Data Using Metrics

Roger Bilisoly

DOI: https://doi.org/10.48550/arXiv.1410.1106

2014-10-05

Applications

Abstract:Researchers have developed ways to generalize the mean and variance to situations in which a data metric is available. We apply the tools developed in Pennec (2006) to categorical data, and show the generality of this approach by considering two quite different applications. First, spelling variability in Middle English is quantified. Second, variability of a finite group (in the sense of group theory) is defined and applied to an example.
Wasserstein Medians: Robustness, PDE Characterization, and Numerics

G. Carlier,E. Chenchene,K. Eichinger

DOI: https://doi.org/10.1137/23m1624786

IF: 2.071

2024-09-21

SIAM Journal on Mathematical Analysis

Abstract:SIAM Journal on Mathematical Analysis, Volume 56, Issue 5, Page 6483-6520, October 2024. We investigate the notion of a Wasserstein median as an alternative to the Wasserstein barycenter, which has become popular but may be sensitive to outliers. In terms of robustness to corrupted data, we indeed show that Wasserstein medians have a breakdown point of approximately [math]. We give explicit constructions of Wasserstein medians in dimension one which enable us to obtain [math] estimates (which do not hold in higher dimensions). We also address dual and multimarginal reformulations. In convex subsets of [math], we connect Wasserstein medians to a minimal (multi)flow problem à la Beckmann and a system of PDEs of Monge–Kantorovich-type, for which we propose a [math]-Laplacian approximation. Our analysis eventually leads to a new numerical method to compute Wasserstein medians, which is based on a Douglas–Rachford scheme applied to the minimal flow formulation of the problem.

mathematics, applied
MaxSkew and MultiSkew: Two R Packages for Detecting, Measuring and Removing Multivariate Skewness

Cinzia Franceschini,Nicola Loperfido

DOI: https://doi.org/10.48550/arXiv.1903.10426

2019-03-26

Abstract:Skewness plays a relevant role in several multivariate statistical techniques. Sometimes it is used to recover data features, as in cluster analysis. In other circumstances, skewness impairs the performances of statistical methods, as in the Hotelling's one-sample test. In both cases, there is the need to check the symmetry of the underlying distribution, either by visual inspection or by formal testing. The R packages MaxSkew and MultiSkew address these issues by measuring, testing and removing skewness from multivariate data. Skewness is assessed by the third multivariate cumulant and its functions. The hypothesis of symmetry is tested either nonparametrically, with the bootstrap, or parametrically, under the normality assumption. Skewness is removed or at least alleviated by projecting the data onto appropriate linear subspaces. Usages of MaxSkew and MultiSkew are illustrated with the Iris dataset.

Computation
Sub-dimensional Mardia measures of multivariate skewness and kurtosis

Joydeep Chowdhury,Subhajit Dutta,Reinaldo B. Arellano-Valle,Marc G. Genton

DOI: https://doi.org/10.1016/j.jmva.2022.105089

IF: 1.387

2022-11-01

Journal of Multivariate Analysis

Abstract:The Mardia measures of multivariate skewness and kurtosis summarize the respective characteristics of a multivariate distribution with two numbers. However, these measures do not reflect the sub-dimensional features of the distribution. Consequently, testing procedures based on these measures may fail to detect skewness or kurtosis present in a sub-dimension of the multivariate distribution. We introduce sub-dimensional Mardia measures of multivariate skewness and kurtosis, and investigate the information they convey about all sub-dimensional distributions of some symmetric and skewed families of multivariate distributions. The maxima of the sub-dimensional Mardia measures of multivariate skewness and kurtosis are considered, as these reflect the maximum skewness and kurtosis present in the distribution, and also allow us to identify the sub-dimension bearing the highest skewness and kurtosis. Asymptotic distributions of the vectors of sub-dimensional Mardia measures of multivariate skewness and kurtosis are derived, based on which testing procedures for the presence of skewness and of deviation from Gaussian kurtosis are developed. The performances of these tests are compared with some existing tests in the literature on simulated and real datasets.

statistics & probability
Symmetrized and non-symmetrized Asymptotic Mean Value Laplacian in metric measure spaces

Andreas Minne,David Tewodrose

2023-10-25

Abstract:The asymptotic mean value Laplacian - AMV Laplacian - extends the Laplace operator from $\mathbb{R}^n$ to metric measure spaces through limits of averaging integrals. The AMV Laplacian is however not a symmetric operator in general. In this paper therefore a symmetric version of the AMV Laplacian is considered, and focus lies on when the symmetric and non-symmetric AMV operators coincide. Besides Riemannian and 3D contact sub-Riemannian manifolds, we show that they are identical on a large class of metric measure spaces including locally Ahlfors regular spaces with vanishing metric-measure boundary. In addition, we study the context of weighted domains of $\mathbb{R}^n$ where the two operators typically differ, and provide concrete formulae for these operators also at points where the weight vanishes.

Analysis of PDEs
Dynamical Mahler Measure: A survey and some recent results

Annie Carter,Matilde Lalín,Michelle Manes,Alison Beth Miller

DOI: https://doi.org/10.48550/arXiv.2204.04101

2022-04-08

Number Theory

Abstract:We study the dynamical Mahler measure of multivariate polynomials and present dynamical analogues of various results from the classical Mahler measure as well as examples of formulas allowing the computation of the dynamical Mahler measure in certain cases. We discuss multivariate analogues of dynamical Kronecker's Lemma and present some improvements on the result for two variables due to Carter, Lal\'in, Manes, Miller, and Mocz.
Symmetric Rearrangement and Geometric Inequalities on Riemannian Manifolds

Richard Stone

2024-11-23

Abstract:This paper starts by introducing results from geometric measure theory to prove symmetric decreasing rearrangement inequalities on $\mathbb{R}^n$, which give multiple proofs of the isoperimetric and Pólya-Szegő inequalities. Then we consider smooth oriented Riemannian manifolds of the form $M^n = (0,\infty)\times \Sigma^{n-1}$, and test what results carry over from the $\mathbb{R}^n$ setting or what assumptions about $M^n$ need to be added. Of particular interest was proving the smooth co-area formula in the Riemannian manifolds setting and re-formulating particular geometric inequalities.

Differential Geometry
Geometric quantile-based measures of multivariate distributional characteristics

Ha-Young Shin,Hee-Seok Oh

2024-07-10

Abstract:Several new geometric quantile-based measures for multivariate dispersion, skewness, kurtosis, and spherical asymmetry are defined. These measures differ from existing measures, which use volumes and are easy to calculate. Some theoretical justification is given, followed by experiments illustrating that they are reasonable measures of these distributional characteristics and computing confidence regions with the desired coverage.

Statistics Theory,Methodology
The Asymptotic Properties of the One-Sample Spatial Rank Methods

Jyrki Möttönen,Klaus Nordhausen,Hannu Oja,Una Radojicic

2023-08-12

Abstract:For a set of $p$-variate data points $\boldsymbol y_1,\ldots,\boldsymbol y_n$, there are several versions of multivariate median and related multivariate sign test proposed and studied in the literature. In this paper we consider the asymptotic properties of the multivariate extension of the Hodges-Lehmann (HL) estimator, the spatial HL-estimator, and the related test statistic. The asymptotic behavior of the spatial HL-estimator and the related test statistic when $n$ tends to infinity are collected, reviewed, and proved, some for the first time though being used already for a longer time. We also derive the limiting behavior of the HL-estimator when both the sample size $n$ and the dimension $p$ tend to infinity.

Statistics Theory
Multifractal Analysis of Measures Arising from Random Substitutions

Andrew Mitchell,Alex Rutar

DOI: https://doi.org/10.1007/s00220-023-04895-3

IF: 2.361

2024-02-26

Communications in Mathematical Physics

Abstract:We study regularity properties of frequency measures arising from random substitutions, which are a generalisation of (deterministic) substitutions where the substituted image of each letter is chosen independently from a fixed finite set. In particular, for a natural class of such measures, we derive a closed-form analytic formula for the -spectrum and prove that the multifractal formalism holds. This provides an interesting new class of measures satisfying the multifractal formalism. More generally, we establish results concerning the -spectrum of a broad class of frequency measures. We introduce a new notion called the inflation word - spectrum of a random substitution and show that this coincides with the -spectrum of the corresponding frequency measure for all . As an application, we obtain closed-form formulas under separation conditions and recover known results for topological and measure theoretic entropy.

physics, mathematical
Derivatives of entropy and the MMSE conjecture

Paul Mansanarez,Guillaume Poly,Yvik Swan

2024-06-16

Abstract:We investigate the entropy $H(\mu,t)$ of a probability measure $\mu$ along the heat flow and more precisely we seek for closed algebraic representations of its derivatives. Provided that $\mu$ admits moments of any order, it is indeed proved in [Guo et al., 2010] that $t\mapsto H(\mu,t)$ is smooth, and in [Ledoux, 2016] that its derivatives at zero can be expressed into multivariate polynomials evaluated in the moments (or cumulants) of $\mu$. In the seminal contribution \cite{Led}, these algebraic expressions are derived through $\Gamma$-calculus techniques which provide implicit recursive formulas for these polynomials. Our main contribution consists in a fine combinatorial analysis of these inductive relations and for the first time to derive closed formulas for the leading coefficients of these polynomials expressions. Building upon these explicit formulas we revisit the so-called "MMSE conjecture" from [Guo et al., 2010] which asserts that two distributions on the real line with the same entropy along the heat flow must coincide up to translation and symmetry. Our approach enables us to provide new conditions on the source distributions ensuring that the MMSE conjecture holds and to refine several criteria proved in [Ledoux, 2016]. As illustrating examples, our findings cover the cases of uniform and Rademacher distributions, for which previous results in the literature were inapplicable.

Information Theory,Probability
Advances on Permutation Multivariate Analysis of Variance for big data

Stefano Bonnini,Getnet Melak Assegie

DOI: https://doi.org/10.2478/stattrans-2022-0022

2022-06-01

Statistics in Transition New Series

Abstract:Abstract In many applications of the multivariate analyses of variance, the classic parametric solutions for testing hypotheses of equality in population means or multisample and multivariate location problems might not be suitable for various reasons. Multivariate multisample location problems lack a comparative study of the power behaviour of the most important combined permutation tests as the number of variables diverges. In particular, it is useful to know under which conditions each of the different tests is preferable in terms of power, how the power of each test increases when the number of variables under the alternative hypothesis diverges, and the power behaviour of each test as the function of the proportion of true alternative hypotheses. The purpose of this paper is to fill the gap in the literature about combined permutation tests, in particular for big data with a large number of variables. A Monte Carlo simulation study was carried out to investigate the power behaviour of the tests, and the application to a real case study was performed to show the utility of the method.
Anti-MANOVA on Compact Manifolds with Applications to 3D Projective Shape Analysis

Hwiyoung Lee,Vic Patrangenaru

DOI: https://doi.org/10.48550/arXiv.1909.00320

2019-09-01

Statistics Theory

Abstract:Methods of hypotheses testing for equality of extrinsic antimeans on compact manifolds are unveiled in this paper. The two and multiple sample problem for antimeans on compact manifolds is addressed for large samples via asymptotic distributions, as well as for small samples using nonparametric bootstrap. An example of face differentiation using 3D VW antimean projective shape analysis for data extracted from digital camera images is also given.
The uniqueness of symmetrizing measure and linear diffusions

Xing Fang,Jiangang Ying,Minzhi Zhao

2009-01-01

Abstract:In this short article, we shall study one-dimensional local Dirichlet spaces. One result, which has its independent interest, is to prove that irreducibility implies the uniqueness of symmetrizing measure for right Markov processes. The other result is to give a representation for any 1-dim local, irreducible and regular Dirichlet space and a necessary and sufficient condition for a Dirichlet space to be regular subspace of another Dirichlet space.

Multivariate medians and measure-symmetrization

Median Shapes

On the Wasserstein Median of Probability Measures

Differentially private multivariate medians

Variance Inequalities for Transformed Fréchet Means in Hadamard Spaces

Sibson's $α$-Mutual Information and its Variational Representations

Bayesian Inference on Multivariate Medians and Quantiles

Generalizing the Mean and Variance to Categorical Data Using Metrics

Wasserstein Medians: Robustness, PDE Characterization, and Numerics

MaxSkew and MultiSkew: Two R Packages for Detecting, Measuring and Removing Multivariate Skewness

Sub-dimensional Mardia measures of multivariate skewness and kurtosis

Symmetrized and non-symmetrized Asymptotic Mean Value Laplacian in metric measure spaces

Dynamical Mahler Measure: A survey and some recent results

Symmetric Rearrangement and Geometric Inequalities on Riemannian Manifolds

Geometric quantile-based measures of multivariate distributional characteristics

The Asymptotic Properties of the One-Sample Spatial Rank Methods

Multifractal Analysis of Measures Arising from Random Substitutions

Derivatives of entropy and the MMSE conjecture

Advances on Permutation Multivariate Analysis of Variance for big data

Anti-MANOVA on Compact Manifolds with Applications to 3D Projective Shape Analysis

The uniqueness of symmetrizing measure and linear diffusions