arXiv subjects

Zakhar Kabluchko

Publications and source records attributed to Zakhar Kabluchko.

At least 127 records · Page 7Linked to original sources

arXiv2017

General Edgeworth expansions with applications to profiles of random trees

We prove an asymptotic Edgeworth expansion for the profiles of certain random trees including binary search trees, random recursive trees and plane-oriented random trees, as the size of the tree goes to infinity. All these models can be seen as special cases of the one-split branching random walk for which we also provide an Edgeworth expansion. These expansions lead to new results on mode, width and occupation numbers of the trees, settling several open problems raised in Devroye and Hwang [Ann. Appl. Probab. 16(2): 886--918, 2006], Fuchs, Hwang and Neininger [Algorithmica, 46 (3--4): 367--407, 2006], and Drmota and Hwang [Adv. in Appl. Probab., 37 (2): 321--341, 2005]. The aforementioned results are special cases and corollaries of a general theorem: an Edgeworth expansion for an arbitrary sequence of random or deterministic functions $\mathbb L_n:\mathbb Z\to\mathbb R$ which converges in the mod-$ϕ$-sense. Applications to Stirling numbers of the first kind will be given in a separate paper.

math.PR↗

arXiv2017

Expected number of real zeros of random Taylor Series

Let $ξ_0,ξ_1,\ldots$ be i.i.d. random variables with zero mean and unit variance. Consider a random Taylor series of the form $f(z)=\sum_{k=0}^\infty ξ_k c_k z^k$, where $c_0,c_1,\ldots$ is a real sequence such that $c_n^2$ is regularly varying with index $γ-1$, where $γ>0$. We prove that $\mathbb{E} N[0,1-ε] \sim \frac{\sqrtγ}{2π} |\log ε|$ as $ε\downarrow 0$, where $N[0,r]$ denotes the number of real zeroes of $f$ in the interval $[0,r]$.

math.PR↗

arXiv2017

High-dimensional limit theorems for random vectors in $\ell_p^n$-balls

In this paper, we prove a multivariate central limit theorem for $\ell_q$-norms of high-dimensional random vectors that are chosen uniformly at random in an $\ell_p^n$-ball. As a consequence, we provide several applications on the intersections of $\ell_p^n$-balls in the flavor of Schechtman and Schmuckenschläger and obtain a central limit theorem for the length of a projection of an $\ell_p^n$-ball onto a line spanned by a random direction $θ\in\mathbb S^{n-1}$. The latter generalizes results obtained for the cube by Paouris, Pivovarov and Zinn and by Kabluchko, Litvak and Zaporozhets. Moreover, we complement our central limit theorems by providing a complete description of the large deviation behavior, which covers fluctuations far beyond the Gaussian scale. In the regime $1\leq p < q$ this displays in speed and rate function deviations of the $q$-norm on an $\ell_p^n$-ball obtained by Schechtman and Zinn, but we obtain explicit constants.

math.FA↗

arXiv2017

A multidimensional analogue of the arcsine law for the number of positive terms in a random walk

Consider a random walk $S_i= ξ_1+\ldots+ξ_i$, $i\in\mathbb N$, whose increments $ξ_1,ξ_2,\ldots$ are independent identically distributed random vectors in $\mathbb R^d$ such that $ξ_1$ has the same law as $-ξ_1$ and $\mathbb P[ξ_1\in H] = 0$ for every affine hyperplane $H\subset \mathbb R^d$. Our main result is the distribution-free formula $$ \mathbb E \left[\sum_{1\leq i_1 < \ldots < i_k\leq n} 1_{\{0\notin \text{conv}(S_{i_1},\ldots, S_{i_k})\}}\right] = 2 \binom n k \frac {B(k, d-1) + B(k, d-3) +\ldots} {2^k k!}, $$ where the $B(k,j)$'s are defined by their generating function $$ (t+1) (t+3) \ldots (t+2k-1) = \sum_{j=0}^{k} B(k,j) t^j. $$ The expected number of $k$-tuples above admits the following geometric interpretation: it is the expected number of $k$-dimensional faces of a randomly and uniformly sampled open Weyl chamber of type $B_n$ that are not intersected by a generic linear subspace $L\subset \mathbb R^n$ of codimension $d$. The case $d=1$ turns out to be equivalent to the classical discrete arcsine law for the number of positive terms in a one-dimensional random walk with continuous symmetric distribution of increments. We also prove similar results for random bridges with no central symmetry assumption required.

math.PR↗

arXiv2017

Convex hulls of random walks: Expected number of faces and face probabilities

Consider a sequence of partial sums $S_i= ξ_1+\dots+ξ_i$, $1\leq i\leq n$, starting at $S_0=0$, whose increments $ξ_1,\dots,ξ_n$ are random vectors in $\mathbb R^d$, $d\leq n$. We are interested in the properties of the convex hull $C_n:=\mathrm{Conv}(S_0,S_1,\dots,S_n)$. Assuming that the tuple $(ξ_1,\dots,ξ_n)$ is exchangeable and a certain general position condition holds, we prove that the expected number of $k$-dimensional faces of $C_n$ is given by the formula $$ \mathbb E [f_k(C_n)] = \frac{2\cdot k!}{n!} \sum_{l=0}^{\infty}\genfrac{[}{]}{0pt}{}{n+1}{d-2l} \genfrac{\{}{\}}{0pt}{}{d-2l}{k+1}, $$ for all $0\leq k \leq d-1$, where $\genfrac{[}{]}{0pt}{}{n}{m}$ and $\genfrac{\{}{\}}{0pt}{}{n}{m}$ are Stirling numbers of the first and second kind, respectively. Further, we compute explicitly the probability that for given indices $0\leq i_1<\dots <i_{k+1}\leq n$, the points $S_{i_1},\dots,S_{i_{k+1}}$ form a $k$-dimensional face of $\mathrm{Conv}(S_0,S_1,\dots,S_n)$. This is done in two different settings: for random walks with symmetrically exchangeable increments and for random bridges with exchangeable increments. These results generalize the classical one-dimensional discrete arcsine law for the position of the maximum due to E. Sparre Andersen. All our formulae are distribution-free, that is do not depend on the distribution of the increments $ξ_k$'s. The main ingredient in the proof is the computation of the probability that the origin is absorbed by a joint convex hull of several random walks and bridges whose increments are invariant with respect to the action of direct product of finitely many reflection groups of types $A_{n-1}$ and $B_n$. This probability, in turn, is related to the number of Weyl chambers of a product-type reflection group that are intersected by a linear subspace in general position.

math.PR↗

arXiv2017

Adaptive Energy Saving Approximation for Random Stationary Processes

We consider a stationary process (with either discrete or continuous time) and find an adaptive approximating stationary process combining approximation quality and supplementary good properties that can be interpreted as additional smoothness or small expense of energy. The problem is solved in terms of the spectral characteristics of the approximated process by using classical analytic methods from prediction theory.

math.PR↗

arXiv2017

Limit theorems for random simplices in high dimensions

Let $r=r(n)$ be a sequence of integers such that $r\leq n$ and let $X_1,\ldots,X_{r+1}$ be independent random points distributed according to the Gaussian, the Beta or the spherical distribution on $\mathbb{R}^n$. Limit theorems for the log-volume and the volume of the random convex hull of $X_1,\ldots,X_{r+1}$ are established in high dimensions, that is, as $r$ and $n$ tend to infinity simultaneously. This includes, Berry-Esseen-type central limit theorems, log-normal limit theorems, moderate and large deviations. Also different types of mod-$ϕ$ convergence are derived. The results heavily depend on the asymptotic growth of $r$ relative to $n$. For example, we prove that the fluctuations of the volume of the simplex are normal (respectively, log-normal) if $r=o(n)$ (respectively, $r\sim αn$ for some $0 < α< 1$).

math.PR↗

arXiv2017

Expected volumes of Gaussian polytopes, external angles, and multiple order statistics

Let $X_1,\ldots,X_n$ be a standard normal sample in $\mathbb R^d$. We compute exactly the expected volume of the Gaussian polytope $\mathrm{conv}[X_1,\ldots,X_n]$, the symmetric Gaussian polytope $\mathrm{conv}[\pm X_1,\ldots,\pm X_n]$, and the Gaussian zonotope $[0,X_1]+\ldots+[0,X_n]$ by exploiting their connection to the regular simplex, the regular crosspolytope, and the cube with the aid of Tsirelson's formula. The expected volumes of these random polytopes are given by essentially the same expressions as the intrinsic volumes and external angles of the regular polytopes. For all these quantities, we obtain asymptotic formulae which are more precise than the results which were known before. More generally, we determine the expected volumes of some heteroscedastic random polytopes including $ \mathrm{conv}[l_1X_1,\ldots,l_nX_n] $ and $ \mathrm{conv}[\pm l_1 X_1,\ldots, \pm l_n X_n], $ where $l_1,\ldots,l_n\geq 0$ are parameters, and the intrinsic volumes of the corresponding deterministic polytopes. Finally, we relate the $k$-th intrinsic volume of the regular simplex $S^{n-1}$ to the expected maximum of independent standard Gaussian random variables $ξ_1,\ldots,ξ_n$ given that the maximum has multiplicity $k$. Namely, we show that $$ V_k(S^{n-1}) = \frac {(2π)^{\frac k2}} {k!} \cdot \lim_{\varepsilon\downarrow 0} \varepsilon^{1-k} \mathbb E [\max\{ξ_1,\ldots,ξ_n\} 1_{\{ξ_{(n)} - ξ_{(n-k+1)}\leq \varepsilon\}}], $$ where $ξ_{(1)} \leq \ldots \leq ξ_{(n)}$ denote the order statistics. A similar result holds for the crosspolytope if we replace $ξ_1,\ldots,ξ_n$ by their absolute values.

math.PR↗

arXiv2017

Convex hulls of random walks, hyperplane arrangements, and Weyl chambers

We give an explicit formula for the probability that the convex hull of an $n$-step random walk in $R^d$ does not contain the origin, under the assumption that the distribution of increments of the walk is centrally symmetric and puts no mass on affine hyperplanes. This extends the formula by Sparre Andersen (1949) for the probability that such random walk in dimension one stays positive. Our result is distribution-free, that is, the probability does not depend on the distribution of increments. This probabilistic problem is shown to be equivalent to either of the two geometric ones: 1) Find the number of Weyl chambers of type $B_n$ intersected by a generic linear subspace of $R^n$ of codimension $d$; 2) Find the conic intrinsic volumes of a Weyl chamber of type $B_n$. We solve the first geometric problem using the theory of hyperplane arrangements. A by-product of our method is a new simple proof of the general formula by Klivans and Swartz (2011) relating the coefficients of the characteristic polynomial of a linear hyperplane arrangement to the conic intrinsic volumes of the chambers constituting its complement. We obtain analogous distribution-free results for Weyl chambers of type $A_{n-1}$ (yielding the probability of absorption of the origin by the convex hull of a generic random walk bridge), type $D_n$, and direct products of Weyl chambers (yielding the absorption probability for the joint convex hull of several random walks or bridges). The simplest case of products of the form $B_1\times \dots \times B_1$ recovers the Wendel formula (1962) for the probability that the convex hull of an i.i.d. multidimensional sample chosen from a centrally symmetric distribution does not contain the origin. We also give an asymptotic analysis of the obtained absorption probabilities as $n \to \infty$, in both cases of fixed and increasing dimension $d$.

math.PR↗

arXiv2017

Monotonicity of expected $f$-vectors for projections of regular polytopes

Let $P_n$ be an $n$-dimensional regular polytope from one of the three infinite series (regular simplices, regular crosspolytopes, and cubes). Project $P_n$ onto a random, uniformly distributed linear subspace of dimension $d\geq 2$. We prove that the expected number of $k$-dimensional faces of the resulting random polytope is an increasing function of $n$. As a corollary, we show that the expected number of $k$-faces of the Gaussian polytope is an increasing function of the number of points used to generate the polytope. Similar results are obtained for the symmetric Gaussian polytope and the Gaussian zonotope.

math.PR↗

arXiv2016

An inclusion-exclusion identity for normal cones of polyhedral sets

For a nonempty polyhedral set $P\subset \mathbb R^d$, let $\mathcal F(P)$ denote the set of faces of $P$, and let $N(P,F)$ be the normal cone of $P$ at the nonempty face $F\in\mathcal F(P)$. We prove that the function $\sum_{F\in\mathcal F(P)}(-1)^{\text{dim} F} 1_{F-N(P,F)}$ equals $1$ if $P$ is bounded, or $0$ if $P$ is unbounded and line-free. Previously, this formula was known to hold everywhere outside some exceptional set of Lebesgue measure $0$ or for polyhedral cones. The case of a not necessarily line-free polyhedral set is also covered by our general theorem.

math.MG↗

arXiv2016

Functional limit theorems for Galton-Watson processes with very active immigration

We prove weak convergence on the Skorokhod space of Galton-Watson processes with immigration, properly normalized, under the assumption that the tail of the immigration distribution has a logarithmic decay. The limits are extremal shot noise processes. By considering marginal distributions, we recover the results of Pakes [Adv. Appl. Probab., 11(1979), 31--62].

math.PR↗

arXiv2016

A leader-election procedure using records

The study of the number of collisions in a Poisson-Dirichlet coalescent leads to the analysis of the following version of a stochastic leader-elec\-tion algorithm. Consider an infinite family of persons, labeled by $1,2,3,\ldots$, who generate iid random numbers from an arbitrary continuous distribution. Those persons who have generated a record value, that is, a value larger than the values of all previous persons, stay in the game, all others must leave. The remaining persons are relabeled by $1,2,3,\ldots$ maintaining their order in the first round, and the election procedure is repeated independently from the past and indefinitely. We prove limit theorems for a number of relevant functionals for this procedure, notably the number of rounds $T(M)$ until all persons among $1,\ldots,M$, except the first one, have left (as $M\to\infty$). For example, we show that the sequence $(T(M)-\log^{*}M)_{M\in\mathbb{N}}$, where $\log^{*}$ denotes the iterated logarithm, is tight, and study its weak subsequential limits. We further provide an appropriate and apparently new kind of normalization (based on tetrations) such that the original labels of persons who stay in the game until round $n$ converge (as $n\to\infty$) to some random non-Poissonian point process and study its properties. The results are applied to study subsequential distributional limits for the number of collisions in the Poisson-Dirichlet coalescent.

math.PR↗

arXiv2016

Leader election using random walks

In the classical leader election procedure all players toss coins independently and those who get tails leave the game, while those who get heads move to the next round where the procedure is repeated. We investigate a generalizion of this procedure in which the labels (positions) of the players who remain in the game are determined using an integer-valued random walk. We study the asymptotics of some relevant quantities for this model such as: the positions of the persons who remained after $n$ rounds; the total number of rounds until all the persons among $1,2,\ldots,M$ leave the game; and the number of players among $1,2,\ldots,M$ who survived the first $n$ rounds. Our results lead to some interesting connection with Galton-Watson branching processes and with the solutions of certain stochastic-fixed point equations arising in the context of the stability of point processes under thinning. We describe the set of solutions to these equations and thus provide a characterization of one-dimensional point processes that are stable with respect to thinning by integer-valued random walks.

math.PR↗

arXiv2016

Some extensions of linear approximation and prediction problems for stationary processes

Let $(B(t))_{t\in Θ}$ with $Θ={\mathbb Z}$ or $Θ={\mathbb R}$ be a wide sense stationary process with discrete or continuous time. The classical linear prediction problem consists of finding an element in $\overline{span}\{B(s),s\le t\}$ providing the best possible mean square approximation to the variable $B(τ)$ with $τ>t$. In this article we investigate this and some other similar problems where, in addition to prediction quality, optimization takes into account other features of the objects we search for. One of the most motivating examples of this kind is an approximation of a stationary process $B$ by a stationary differentiable process $X$ taking into account the kinetic energy that $X$ spends in its approximation efforts.

math.PR↗

arXiv2016

Ergodic decompositions of stationary max-stable processes in terms of their spectral functions

We revisit conservative/dissipative and positive/null decompositions of stationary max-stable processes. Originally, both decompositions were defined in an abstract way based on the underlying non-singular flow representation. We provide simple criteria which allow to tell whether a given spectral function belongs to the conservative/dissipative or positive/null part of the de Haan spectral representation. Specifically, we prove that a spectral function is null-recurrent iff it converges to $0$ in the Cesàro sense. For processes with locally bounded sample paths we show that a spectral function is dissipative iff it converges to $0$. Surprisingly, for such processes a spectral function is integrable a.s. iff it converges to $0$ a.s. Based on these results, we provide new criteria for ergodicity, mixing, and existence of a mixed moving maximum representation of a stationary max-stable process in terms of its spectral functions. In particular, we study a decomposition of max-stable processes which characterizes the mixing property.

math.PR↗

arXiv2016

Mode and Edgeworth expansion for the Ewens distribution and the Stirling numbers

We provide asymptotic expansions for the Stirling numbers of the first kind and, more generally, the Ewens (or Karamata-Stirling) distribution. Based on these expansions, we obtain some new results on the asymptotic properties of the mode and the maximum of the Stirling numbers and the Ewens distribution. For arbitrary $θ>0$ and for all sufficiently large $n\in\mathbb N$, the unique maximum of the Ewens probability mass function $$ \mathbb L_n(k) = \frac{θ^k}{θ(θ+1)\ldots(θ+n-1)} \genfrac{[}{]}{0pt}{}{n}{k}, \quad k=1,\ldots,n, $$ is attained at $k= \left\lfloor θ\log n + \frac{θΓ'(θ)}{Γ(θ)} - \frac 12\right\rfloor$ or $k=\left\lceil θ\log n + \frac{θΓ'(θ)}{Γ(θ)} + \frac 12\right\rceil$. We prove that the mode is $$ k=\left\lfloor θ\log n - \frac{θΓ'(θ)}{Γ(θ)}\right\rfloor $$ for a set of $n$'s of asymptotic density $1$, yet this formula is not true for infinitely many $n$'s.

math.PR↗

arXiv2016

Edgeworth expansions for profiles of lattice branching random walks

Consider a branching random walk on $\mathbb Z$ in discrete time. Denote by $L_n(k)$ the number of particles at site $k\in\mathbb Z$ at time $n\in\mathbb N_0$. By the profile of the branching random walk (at time $n$) we mean the function $k\mapsto L_n(k)$. We establish the following asymptotic expansion of $L_n(k)$, as $n\to\infty$: $$ e^{-φ(0)n} L_n(k) = \frac{e^{-\frac 12 x_n^2(k)}}{\sqrt {2πφ''(0) n}} \sum_{j=0}^r \frac{F_j(x_n(k))}{n^{j/2}} + o\left(n^{-\frac{r+1}{2}}\right) \quad a.s., $$ where $r\in\mathbb N_0$ is arbitrary, $φ(β)=\log \sum_{k\in\mathbb Z} e^{βk} \mathbb E L_1(k)$ is the cumulant generating function of the intensity of the branching random walk and $$ x_n(k) = \frac{k-φ'(0) n}{\sqrt{φ''(0)n}}. $$ The expansion is valid uniformly in $k\in\mathbb Z$ with probability $1$ and the $F_j$'s are polynomials whose random coefficients can be expressed through the derivatives of $φ$ and the derivatives of the limit of the Biggins martingale at $0$. Using exponential tilting, we also establish more general expansions covering the whole range of the branching random walk except its extreme values. As an application of this expansion for $r=0,1,2$ we recover in a unified way a number of known results and establish several new limit theorems. In particular, we study the a.s. behavior of the individual occupation numbers $L_n(k_n)$, where $k_n\in\mathbb Z$ depends on $n$ in some regular way. We also prove a.s. limit theorems for the mode $\arg \max_{k\in\mathbb Z} L_n(k)$ and the height $\max_{k\in\mathbb Z} L_n(k)$ of the profile. The asymptotic behavior of these quantities depends on whether the drift parameter $φ'(0)$ is integer, non-integer rational, or irrational.

math.PR↗