Showing 1–14 of 14 results for author: Fresen, D

Optimal tail comparison under convex majorization

Authors: Daniel J. Fresen

Abstract: Following results of Kemperman and Pinelis, we show that if $X$ and $Y$ are real valued random variables such that $\mathbb{E}\left\vert Y\right\vert<\infty$ and for all non-decreasing convex $\varphi:\mathbb{R}\rightarrow [0,\infty)$, $\mathbb{E}\varphi(X)\leq\mathbb{E}\varphi(Y)$, then for all $s\in\mathbb{R}$ with $\mathbb{P}\left\{Y>s\right\}\neq 0$,… ▽ More Following results of Kemperman and Pinelis, we show that if $X$ and $Y$ are real valued random variables such that $\mathbb{E}\left\vert Y\right\vert<\infty$ and for all non-decreasing convex $\varphi:\mathbb{R}\rightarrow [0,\infty)$, $\mathbb{E}\varphi(X)\leq\mathbb{E}\varphi(Y)$, then for all $s\in\mathbb{R}$ with $\mathbb{P}\left\{Y>s\right\}\neq 0$, $\mathbb{P}\left\{X\geq\mathbb{E}\left(Y:Y>s\right)\right\}\leq\mathbb{P}\left\{Y>s\right\}$. This bound is sharp in essentially the strictest possible sense: for any such $Y$ and $s$ there exists such an $X$ with $\mathbb{P}\left\{X\geq \mathbb{E}\left(Y:Y>s\right)\right\}=\mathbb{P}\left\{Y>s\right\}$. △ Less

Submitted 5 July, 2022; originally announced July 2022.

Comments: 9 pages. Most of the material here was originally part of arXiv:2203.12523 and/or arXiv:1812.10938 and now stands as a paper on its own

MSC Class: 60E05; 60E15

Moment free deviation inequalities for linear combinations of independent random variables with power-type tails

Authors: Daniel J. Fresen

Abstract: We present order of magnitude estimates for the quantiles of non-negative linear combinations of non-negative random variables, as well as deviation inequalities for general linear combinations of independent random variables, under the assumption that all random variables satisfy the same power-type tail bound on $\mathbb{P}\{\left\vert X_i\right\vert>t\}$ of the form $t^{-q}$, $t^{-q/2}$ or… ▽ More We present order of magnitude estimates for the quantiles of non-negative linear combinations of non-negative random variables, as well as deviation inequalities for general linear combinations of independent random variables, under the assumption that all random variables satisfy the same power-type tail bound on $\mathbb{P}\{\left\vert X_i\right\vert>t\}$ of the form $t^{-q}$, $t^{-q/2}$ or $t^{-q/2}(\ln t)^{q/2}$, for $q>2$. The third type is applicable in the nonlinear setting. In the situations we consider, these results improve on classical estimates of Nagaev. △ Less

Submitted 15 August, 2022; v1 submitted 5 July, 2022; originally announced July 2022.

Comments: 29 pages. A significant amount of the material here was originally part of arXiv:2203.12523 and/or arXiv:1812.10938 and now stands as a paper on its own. This paper used to have the title 'Variations and extensions of the Gaussian concentration inequality, Part IIa: Linear case'

MSC Class: 60E05; 60E15; 46N30

Convex subsets of non-convex Lorentz balls

Authors: Daniel J. Fresen

Abstract: Many star bodies have convex subsets with approximately the same Gaussian measure (of the complement). Inspired by this phenomenon, and in connection with the randomized Dvoretzky theorem for Lorentz spaces, we derive bounds on the distribution of certain functions of a Gaussian random vector by approximating their sub-level sets by convex subsets. Many star bodies have convex subsets with approximately the same Gaussian measure (of the complement). Inspired by this phenomenon, and in connection with the randomized Dvoretzky theorem for Lorentz spaces, we derive bounds on the distribution of certain functions of a Gaussian random vector by approximating their sub-level sets by convex subsets. △ Less

Submitted 21 June, 2022; originally announced June 2022.

Comments: 18 pages. Most of the material here was originally part of arXiv:2104.11974 and now stands as a paper on its own

MSC Class: 52A27; 52A30; 60E05; 60E15

Variations and extensions of the Gaussian concentration inequality, Part II

Authors: Daniel J. Fresen

Abstract: We prove concentration inequalities for $f\left( X\right) $ about its median, where $X$ is a random vector in $\mathbb{R}^n$ with independent heavy tailed coordinates of Weibull or power type, and $f:\mathbb{R}^n\rightarrow\mathbb{R}$ is a locally Lipschitz function. This paper is part of a series of four papers, Part I, Part II and two supporting papers. It can be read independently of Part I. We prove concentration inequalities for $f\left( X\right) $ about its median, where $X$ is a random vector in $\mathbb{R}^n$ with independent heavy tailed coordinates of Weibull or power type, and $f:\mathbb{R}^n\rightarrow\mathbb{R}$ is a locally Lipschitz function. This paper is part of a series of four papers, Part I, Part II and two supporting papers. It can be read independently of Part I. △ Less

Submitted 10 August, 2022; v1 submitted 23 March, 2022; originally announced March 2022.

Comments: 14 pages. The original paper, which was uploaded as a separate submission (arXiv:1812.10938), has been split into two papers, Part I and Part II; this is Part II

MSC Class: 60E05; 60E15; 46N30

Random Euclidean embeddings in finite dimensional Lorentz spaces

Authors: Daniel J. Fresen

Abstract: Quantitative bounds for random embeddings of $\mathbb{R}^{k}$ into Lorentz sequence spaces are given, with improved dependence on $\varepsilon$. Quantitative bounds for random embeddings of $\mathbb{R}^{k}$ into Lorentz sequence spaces are given, with improved dependence on $\varepsilon$. △ Less

Submitted 24 April, 2021; originally announced April 2021.

Comments: 32 pages

MSC Class: 46B06; 46B07; 46B09; 52A21; 52A23

A simplified proof of CLT for convex bodies

Authors: Daniel J. Fresen

Abstract: We present a short proof of Klartag's central limit theorem for convex bodies, using only the most classical facts about log-concave functions. An appendix is included where we give the proof that thin shell implies CLT. The paper is accessible to anyone. We present a short proof of Klartag's central limit theorem for convex bodies, using only the most classical facts about log-concave functions. An appendix is included where we give the proof that thin shell implies CLT. The paper is accessible to anyone. △ Less

Submitted 19 July, 2019; v1 submitted 15 July, 2019; originally announced July 2019.

Comments: 11 pages, minor changes

MSC Class: 52A20; 52A23; 60F05

Variations and extensions of the Gaussian concentration inequality, Part I

Authors: Daniel J. Fresen

Abstract: The classical Gaussian concentration inequality for Lipschitz functions is adapted to a setting where the classical assumptions (i.e. Lipschitz and Gaussian) are not met. The theory is more direct than much of the existing theory designed to handle related generalizations. An application is presented to linear combinations of heavy tailed random variables. The classical Gaussian concentration inequality for Lipschitz functions is adapted to a setting where the classical assumptions (i.e. Lipschitz and Gaussian) are not met. The theory is more direct than much of the existing theory designed to handle related generalizations. An application is presented to linear combinations of heavy tailed random variables. △ Less

Submitted 13 May, 2022; v1 submitted 28 December, 2018; originally announced December 2018.

Comments: 18 pages. The original arXiv version has been split into two separate papers, Part I and Part II; this is Part I. To appear in Quaest. Math

MSC Class: 60E05; 60E15; 52A27; 52A30

Concentration of random polytopes around the expected convex hull

Authors: Daniel J. Fresen, Richard A. Vitale

Abstract: We provide a streamlined proof and improved estimates for the weak multivariate Gnedenko law of large numbers on concentration of random polytopes within the space of convex bodies (in a fixed or a high dimensional setting), as well as a corresponding strong law of large numbers. We provide a streamlined proof and improved estimates for the weak multivariate Gnedenko law of large numbers on concentration of random polytopes within the space of convex bodies (in a fixed or a high dimensional setting), as well as a corresponding strong law of large numbers. △ Less

Submitted 9 March, 2014; v1 submitted 11 February, 2014; originally announced February 2014.

Comments: 9 pages

MSC Class: 60D05; 60F99; 52A20; 52A22; 52B11

Explicit Euclidean Embeddings in Permutation Invariant Normed Spaces

Authors: Daniel Fresen

Abstract: Let $(X,\left\Vert \cdot \right\Vert )$ be a real normed space of dimension $N\in \mathbb{N}$ with a basis $(e_{i})_{1}^{N}$ such that the norm is invariant under coordinate permutations. Assume for simplicity that the basis constant is at most $2$. Consider any $n\in \mathbb{N}$ and $0<\varepsilon <1/4$ such that $n\leq c(\log \varepsilon ^{-1})^{-1}\log N$. We provide an explicit construction of… ▽ More Let $(X,\left\Vert \cdot \right\Vert )$ be a real normed space of dimension $N\in \mathbb{N}$ with a basis $(e_{i})_{1}^{N}$ such that the norm is invariant under coordinate permutations. Assume for simplicity that the basis constant is at most $2$. Consider any $n\in \mathbb{N}$ and $0<\varepsilon <1/4$ such that $n\leq c(\log \varepsilon ^{-1})^{-1}\log N$. We provide an explicit construction of a matrix that generates a $(1+\varepsilon )$ embedding of $\ell _{2}^{n}$ into $X$. △ Less

Submitted 31 December, 2013; originally announced January 2014.

Comments: 14 pages

MSC Class: 46B06; 46B07; 52A20; 52A21; 52A23

Euclidean arrangements in Banach spaces

Authors: Daniel J. Fresen

Abstract: We study the way in which the Euclidean subspaces of a Banach space fit together, somewhat in the spirit of the Kašin decomposition. The main tool that we introduce is an estimate regarding the convex hull of a convex body in John's position with a Euclidean ball of a given radius, which leads to a new and simplified proof of the randomized isomorphic Dvoretzky theorem. Our results also include a… ▽ More We study the way in which the Euclidean subspaces of a Banach space fit together, somewhat in the spirit of the Kašin decomposition. The main tool that we introduce is an estimate regarding the convex hull of a convex body in John's position with a Euclidean ball of a given radius, which leads to a new and simplified proof of the randomized isomorphic Dvoretzky theorem. Our results also include a characterization of spaces with nontrivial cotype in terms of arrangements of Euclidean subspaces. △ Less

Submitted 4 May, 2015; v1 submitted 21 September, 2013; originally announced September 2013.

Comments: 22 pages

MSC Class: 46B20; 52A23; 46B09; 52A21; 46B07

Exponential and Gaussian behavior in the tails of multivariate functions

Authors: Daniel Fresen

Abstract: We observe that approximate copies of the function $Λ_{n}:\mathbb{R}^{n}\rightarrow (0,\infty )$ defined by \begin{equation*} Λ_{n}(x)=\exp \left( -x_{1}-π\sum_{i=2}^{n}x_{i}^{2}\right) \end{equation*} appear in the tails of a large class of functions, with properties related to coordinate independence, convexity, homotheticity, and homogeneity. The function $Λ_{n}$ is an entropy maximizer (on a h… ▽ More We observe that approximate copies of the function $Λ_{n}:\mathbb{R}^{n}\rightarrow (0,\infty )$ defined by \begin{equation*} Λ_{n}(x)=\exp \left( -x_{1}-π\sum_{i=2}^{n}x_{i}^{2}\right) \end{equation*} appear in the tails of a large class of functions, with properties related to coordinate independence, convexity, homotheticity, and homogeneity. The function $Λ_{n}$ is an entropy maximizer (on a half-space) that is uniquely determined by a homogeneity condition together with rotational invariance about the $x_{1}$ direction and its behavior near the origin. These results are connected to the limiting Poisson point processes found near the edges of large random samples, as well as the conditioning of random vectors on certain rare events, and can be thought of as variations of Laplace's method for estimating integrals. △ Less

Submitted 18 June, 2019; v1 submitted 25 August, 2011; originally announced August 2011.

Comments: 28 pages

MSC Class: 26B35; 60F05; 60G55; 60G70

Comments on the floating body and the hyperplane conjecture

Authors: Daniel Fresen

Abstract: We provide a reformulation of the hyperplane conjecture (the slicing problem) in terms of the floating body and give upper and lower bounds on the logarithmic Hausdorff distance between an arbitrary convex body $K\subset \mathbb{R}^{d}$\ and the convex floating body $K_δ$ inside $K$. We provide a reformulation of the hyperplane conjecture (the slicing problem) in terms of the floating body and give upper and lower bounds on the logarithmic Hausdorff distance between an arbitrary convex body $K\subset \mathbb{R}^{d}$\ and the convex floating body $K_δ$ inside $K$. △ Less

Submitted 20 February, 2011; v1 submitted 13 February, 2011; originally announced February 2011.

Comments: 8 pages

MSC Class: 52A23; 52A20; 52A21; 52A38

Simultaneous concentration of order statistics

Authors: Daniel Fresen

Abstract: Let $μ$ be a probability measure on $\mathbb{R}$ with cumulative distribution function $F$, $(x_{i})_{1}^{n}$ a large i.i.d. sample from $μ$, and $F_{n}$ the associated empirical distribution function. The Glivenko-Cantelli theorem states that with probability 1, $F_{n}$ converges uniformly to $F$. In so doing it describes the macroscopic structure of $\{x_{i}\}_{1}^{n}$, however it is insensitive… ▽ More Let $μ$ be a probability measure on $\mathbb{R}$ with cumulative distribution function $F$, $(x_{i})_{1}^{n}$ a large i.i.d. sample from $μ$, and $F_{n}$ the associated empirical distribution function. The Glivenko-Cantelli theorem states that with probability 1, $F_{n}$ converges uniformly to $F$. In so doing it describes the macroscopic structure of $\{x_{i}\}_{1}^{n}$, however it is insensitive to the position of individual points. Indeed any subset of $o(n)$ points can be perturbed at will without disturbing the convergence. We provide several refinements of the Glivenko-Cantelli theorem which are sensitive not only to the global structure of the sample but also to individual points. Our main result provides conditions that guarantee simultaneous concentration of all order statistics. The example of main interest is the normal distribution. △ Less

Submitted 20 February, 2011; v1 submitted 6 February, 2011; originally announced February 2011.

Comments: 10 pages

MSC Class: 62G30; 60G55

A multivariate Gnedenko law of large numbers

Authors: Daniel Fresen

Abstract: We show that the convex hull of a large i.i.d. sample from an absolutely continuous log-concave distribution approximates a predetermined convex body in the logarithmic Hausdorff distance and in the Banach-Mazur distance. For log-concave distributions that decay super-exponentially, we also have approximation in the Hausdorff distance. These results are multivariate versions of the Gnedenko law of… ▽ More We show that the convex hull of a large i.i.d. sample from an absolutely continuous log-concave distribution approximates a predetermined convex body in the logarithmic Hausdorff distance and in the Banach-Mazur distance. For log-concave distributions that decay super-exponentially, we also have approximation in the Hausdorff distance. These results are multivariate versions of the Gnedenko law of large numbers, which guarantees concentration of the maximum and minimum in the one-dimensional case. We provide quantitative bounds in terms of the number of points and the dimension of the ambient space. △ Less

Submitted 21 October, 2013; v1 submitted 25 January, 2011; originally announced January 2011.

Comments: Published in at http://dx.doi.org/10.1214/12-AOP804 the Annals of Probability (http://www.imstat.org/aop/) by the Institute of Mathematical Statistics (http://www.imstat.org)

Report number: IMS-AOP-AOP804

Journal ref: Annals of Probability 2013, Vol. 41, No. 5, 3051-3080