Source author record

Milan Merkle

Milan Merkle appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

math.CA math.ST Statistics Theory Computational Geometry Data Structures and Algorithms math.PR

Catalog footprint

What is connected

7works

6topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2020arXiv

A tight Hermite-Hadamard's inequality and a generic method for comparison between residuals of inequalities with convex functions

We present a tight parametrical Hermite-Hadamard type inequality with probability measure, which yields a considerably closer upper bound for the mean value of convex function than the classical one. Our inequality becomes equality not only with affine functions, but also with a family of V-shaped curves determined by the parameter. The residual (error) of this inequality is strictly smaller than in the classical Hermite-Hadamard inequality under any probability measure and with all non-affine convex functions. In the framework of Karamata's theorem on the inequalities with convex functions, we propose a method of measuring a global performance of inequalities in terms of average residuals over functions of the type $x\mapsto |x-u|$. Using average residuals enables comparing two or more inequalities as themselves, with same or different measures and without referring to a particular function. Our method is applicable to all Karamata's type inequalities, with integrals or sums. A numerical experiment with three different measures indicates that the average residual in our inequality is about 4 times smaller than in classical right Hermite-Hadamard, and also is smaller than in Jensen's inequality, with all three measures.

preprint2020arXiv

On bounds for the mode and median of the generalized hyperbolic and related distributions

Except for certain parameter values, a closed form formula for the mode of the generalized hyperbolic (GH) distribution is not available. In this paper, we exploit results from the literature on modified Bessel functions and their ratios to obtain simple but tight two-sided inequalities for the mode of the GH distribution for general parameter values. As a special case, we deduce tight two-sided inequalities for the mode of the variance-gamma (VG) distribution, and through a similar approach we also obtain tight two-sided inequalities for the mode of the McKay Type I distribution. The analogous problem for the median is more challenging, but we conjecture some monotonicity results for the median of the VG and McKay Type I distributions, from we which we conjecture some tight two-sided inequalities for their medians. Numerical experiments support these conjectures and also lead us to a conjectured tight lower bound for the median of the GH distribution.

preprint2016arXiv

ABCDepth: efficient algorithm for Tukey depth

We present a new algorithm for Tukey (halfspace) depth level sets and its implementation. Given $d$-dimensional data set for any $d\geq 2$, the algorithm is based on representation of level sets as intersections of balls in $R^d$, and can be easily adapted to related depths (Type D, Zuo and Serfling (Ann. Stat. {\bf 28} (2000), 461--482)). The algorithm complexity is $O(dn^2 + n^2\log{n})$ where $n$ is the data set size. Examples with real and synthetic data show that the algorithm is much faster than other implemented algorithms and that it can accept thousands of multidimensional observations, while other algorithms are tested with two-dimensional data or with a couple of hundreds multidimensional observations.

preprint2015arXiv

An axiomatic integral and a multivariate mean value theorem

In order to investigate minimal sufficient conditions for an abstract integral to belong to the convex hull of the integrand, we propose a system of axioms under which it happens. If the integrand is a continuous $R^n$-valued function over a path connected topological space, we prove that any such integral can be represented as a convex combination of values of the integrand in at most $n$ points, which yields an ultimate multivariate mean value theorem.

preprint2012arXiv

Completely monotone functions - a digest

This work has a purpose to collect selected facts about the completely monotone (CM) functions that can be found in books and papers devoted to different areas of mathematics. We opted for lesser known ones, and for those which may help determining whether or not a given function is completely monotone. In particular, we emphasize the role of representation of a CM function as the Laplace transform of a measure, and we present and discuss a little known connection with log-convexity. Some of presented methods are illustrated by several examples involving Gamma and related functions.

preprint2011arXiv

Young's functional with Lebesgue-Stieltjes integrals

For non-decreasing real functions $f$ and $g$, we consider the functional $ T(f,g ; I,J)=\int_{I} f(x)\di g(x) + \int_J g(x)\di f(x)$, where $I$ and $J$ are intervals with $J\subseteq I$. In particular case with $I=[a,t]$, $J=[a,s]$, $s\leq t$ and $g(x)=x$, this reduces to the expression in classical Young's inequality. We survey some properties of Lebesgue-Stieltjes interals and present a new simple proof for change of variables. Further, we formulate a version of Young's inequality with respect to arbitrary positive finite measure on real line including a purely discrete case, and discuss an application related to medians of probability distributions and a summation formula that involves values of a function and its inverse at integers.

preprint2008arXiv

Spatial medians, depth functions and multivariate Jensen's inequality

For any given partial order in a $d$-dimensional euclidean space, under mild regularity assumptions, we show that the intersection of closed (generalized) intervals containing more than 1/2 of the probability mass, is a non-empty compact interval. This property is shared with common intervals on real line, where the intersection is the median set of the underlying probability distribution. So obtained multivariate medians with respect to a partial order, can be observed as special cases of centers of distribution in the sense of type D depth functions introduced by Y. Zuo and R. Serfling, {\em Ann. Stat.}, {\bf 28} (2000), 461-482. We show that the halfspace depth function can be realized via compact convex sets, or, for example, closed balls, in place of halfspaces, and discuss structural properties of halfspace and related depth functions and their centers. Among other things, we prove that, in general, the maximal guaranteed depth is $\frac{1}{d+1}$. As an application of these results, we provide a Jensen's type inequality for functions of several variables, with medians in place of expectations, which is an extension of the previous work by M. Merkle, {\em Stat. Prob. Letters}, {\bf 71} (2005), 277--281.}