Integer-valued Random Variables Research Articles

On Poisson Approximation

AbstractThe problem of evaluating the accuracy of Poisson approximation to the distribution of a sum of independent integer-valued random variables has attracted a lot of attention in the past six decades. Among authors who contributed to the topic are Prokhorov, Kolmogorov, LeCam, Shorgin, Barbour, Hall, Deheuvels, Pfeifer, Roos, and many others. From a practical point of view, the problem has important applications in insurance, reliability theory, extreme value theory, etc. From a theoretical point of view, the topic provides insights into Kolmogorov’s problem concerning the accuracy of approximation of the distribution of a sum of independent random variables by infinitely divisible laws. The task of establishing an estimate of the accuracy of Poisson approximation with a correct (the best possible) constant at the leading term remained open for decades. We present a solution to that problem in the case where the accuracy of approximation is evaluated in terms of the point metric. We generalise and sharpen the corresponding inequalities established by preceding authors. A new result is established for the intensively studied topic of compound Poisson approximation to the distribution of a sum of integer-valued r.v.s.

Bernoulli sums and Rényi entropy inequalities

We investigate the Rényi entropy of sums of independent integer-valued random variables through Fourier theoretic means, and give sharp comparisons between the variance and the Rényi entropy for sums of independent Bernoulli random variables. As applications, we prove that a discrete “min-entropy power” is superadditive with respect to convolution modulo a universal constant, and give new bounds on an entropic generalization of the Littlewood-Offord problem that are sharp in the “Poisson regime”.

Local limit theorems without assuming finite third moment

One of the most fundamental probabilities is the probability at a particular point. The local limit theorem is the well-known theorem that estimates this probability. In this paper, we estimate this probability by the density function of normal distribution in the case of lattice integer-valued random variables. Our technique is the characteristic function method. We complete to relax the third moment condition of Siripraparat and Neammanee (J. Inequal. Appl. 2021:57, 2021) and the references therein and also obtain explicit constants of the error bound.

Edgeworth expansions for independent bounded integer valued random variables

We obtain asymptotic expansions for local probabilities of partial sums for uniformly bounded independent but not necessarily identically distributed integer-valued random variables. The expansions involve products of polynomials and trigonometric polynomials. These results also have counterparts for triangular arrays. Our results do not require any additional assumptions. As an application of our expansions we find necessary and sufficient conditions for the classical Edgeworth expansion. It turns out that there are two possible obstructions for the validity of the Edgeworth expansion of order r. First, the distance between the distribution of the underlying partial sums modulo some h∈N and the uniform distribution could fail to be o(σN1−r), where σN is the standard deviation of the partial sum. Second, this distribution could have the required closeness but this closeness is unstable, in the sense that it could be destroyed by removing finitely many terms. In the first case, the expansion of order r fails. In the second case it may or may not hold depending on the behavior of the derivatives of the characteristic functions of the summands whose removal causes the break-up of the uniform distribution. We also show that a quantitative version of the classical Prokhorov condition (for the strong local central limit theorem) is sufficient for Edgeworth expansions, and moreover this condition is, in some sense, optimal.

On some bivariate discrete distributions with multivariate components

Consider a random vector (X,Y) where X=(X1,X2, ...., Xs,) and Y=(Y1,Y2, ...., Ys) with Xi, Yi, i=1,2,..., s independent non-negative, integer-valued random variables with finite support and such that X>=Y. We show that in the case where the distribution of (YlX=n) is of a certain structural form then there exists a relationship between the distributions of Y and of Y|(X=Y) which uniquely determines the distribution of X. The relationship in question is less stringent that the condition of independence between Y and X-Y usually involved in pro¬blems of this nature. Examples are given to illustrate the result. The case where X,Y have infinite support has been examined earlier by the author.

On Strong Deviation Theorems Concerning Arrays of Non-Negative Integer-Valued Random Variables

On Strong Deviation Theorems Concerning Arrays of Non-Negative Integer-Valued Random Variables

On the large O−rates of convergence in limit theorems for compound random sums of arrays of row-wise independent random variables

Compound random sums are extensions of random sums when the random number of summands is a partial sum of independent and identically distributed positive integer-valued random variables, which assumed independent of summands. In the paper, upper bounds for the large O?rates of convergence in weak limit theorems for compound random sums of arrays of row-wise independent random variables, in term of Trotter distance are studied. The main results are approximation theorems which give the Trotter distance between normalized compound random sums of the given independent random variables and the compound ??decomposable random variables. By these results the converging rates in central limit theorem, weak law of large numbers and stable limit theorem for compound random sums are then established. The obtained results in this paper are closely related to the classical ones.

On the accuracy of the poissonisation in the infinite occupancy scheme

We obtain asymptotic accuracy of the poissonisation in the infinite occupancy scheme. All results are obtained for integer-valued random variables having a regularly varying distribution.

A Note on the Accuracy of Normal Approximation of Random Quantities

For a sequence of independent, identically distributed random variable (iid rv's) [Formula: see text] and a sequence of integer-valued random variables [Formula: see text], define the random quantiles as [Formula: see text], where [Formula: see text] denote the largest integer less than or equal to [Formula: see text], and [Formula: see text] the [Formula: see text]th order statistic in a sample [Formula: see text] and [Formula: see text]. In this note, the limiting distribution and its exact order approximation are obtained for [Formula: see text]. The limiting distribution result we obtain extends the work of several including Wretman[Formula: see text]. The exact order of normal approximation generalizes the fixed sample size results of Reiss[Formula: see text]. AMS 2000 subject classification: 60F12; 60F05; 62G30.

Random Gap Processes and Asymptotically Complete Sequences

We study a process of generating random positive integer weight sequences \(\{ W_n \}\) where the gaps between the weights \(\{ X_n = W_n - W_{n-1} \}\) are i.i.d. positive integer-valued random variables. The main result of the paper is that if the gap distribution has a moment generating function with large enough radius of convergence, then the weight sequence is almost surely asymptotically m-complete for every \(m\ge 2\), i.e. every large enough multiple of the greatest common divisor (gcd) of gap values can be written as a sum of m distinct weights for any fixed \(m \ge 2\). Under the weaker assumption of finite \(\frac{1}{2}\)-moment for the gap distribution, we also show the simpler result that, almost surely, the resulting weight sequence is asymptotically complete, i.e. all large enough multiples of the gcd of the possible gap values can be written as a sum of distinct weights.

Poisson approximation in terms of the Gini–Kantorovich distance

It is long known that the distribution of a sum Sn of independent non-negative integer-valued random variables can often be approximated by a Poisson law: Sn≈πλ, where . The problem of evaluating the accuracy of such approximation has attracted a lot of attention in the past six decades. From a practical point of view, the problem has important applications in insurance, reliability theory, extreme value theory, etc.; from a theoretical point of view, it provides insights into Kolmogorov’s problem. Among popular metrics considered in the literature is the Gini–Kantorovich distance dG. The task of establishing an estimate of dG(Sn;πλ) with correct (the best possible) constant at the leading term remained open for a long while. The paper presents a solution to that problem. A first-order asymptotic expansion is established as well. We show that the accuracy of approximation can be considerably better if the random variables obey an extra condition involving the first two moments. A sharp estimate of the accuracy of shifted (translated) Poisson approximation is established as well.

On the discrepancy of random subsequences of $\{n\alpha \}$, II

Let $\alpha$ be an irrational number, let $X_1, X_2, \ldots$ be independent, identically distributed, integer-valued random variables, and put $S_k=\sum_{j=1}^k X_j$. Assuming that $X_1$ has finite variance or heavy tails $P (|X_1|>t)\sim ct^{-\beta}$, $0<\beta<2$, in Part I of this paper we proved that, up to logarithmic factors, the order of magnitude of the discrepancy $D_N (S_k \alpha)$ of the first $N$ terms of the sequence $\{S_k \alpha\}$ is $O(N^{-\tau})$, where $\tau= \min (1/(\beta \gamma), 1/2)$ (with $\beta=2$ in the case of finite variances) and $\gamma$ is the strong Diophantine type of $\alpha$. This shows a change of behavior of the discrepancy at $\beta\gamma=2$. In this paper we determine the exact order of magnitude of $D_N (S_k \alpha)$ for $\beta\gamma<1$, and determine the limit distribution of $N^{-1/2} D_N (S_k \alpha)$. We also prove a functional version of these results describing the asymptotic behavior of a wide class of functionals of the sequence $\{S_k \alpha\}$. Finally, we extend our results to the discrepancy of $\{S_k\}$ for general random walks $S_k$ without arithmetic conditions on $X_1$, assuming only a mild polynomial rate on the weak convergence of $\{S_k\}$ to the uniform distribution.

Poisson approximation with applications to stochastic geometry

This article compares the distributions of integer-valued random variables and Poisson random variables. It considers the total variation and the Wasserstein distance and provides, in particular, explicit bounds on the pointwise difference between the cumulative distribution functions. Special attention is dedicated to estimating the difference when the cumulative distribution functions are evaluated at 0. This permits to approximate the minimum (or maximum) of a collection of random variables by a suitable random variable in the Kolmogorov distance. The main theoretical results are obtained by combining the Chen-Stein method with size-bias coupling and a generalization of size-bias coupling for integer-valued random variables developed herein. A wide variety of applications are then discussed with a focus on stochastic geometry. In particular, transforms of the minimal circumscribed radius and the maximal inradius of Poisson-Voronoi tessellations as well as the minimal inter-point distance of the points of a Poisson process are considered and bounds for their Kolmogorov distances to extreme value distributions are derived.

Discrete line integral on uniform grids: Probabilistic interpretation and applications

Following the methodology developed by (Comput. Math. Appl. 33 (1997) 81–104), we define a discrete version of gradient vector and associated line integral along arbitrary path connecting two nodes of uniform grid. An exponential representation of joint survival function of bivariate discrete non-negative integer-valued random variables in terms of discrete line integral is established. We apply it to generate a discrete analogue of the Sibuya-type aging property, incorporating many classical and new bivariate discrete models. Several characterizations and closure properties of this class of bivariate discrete distributions are presented.

Computing probabilities of integer-valued random variables by recurrence relations

We derive a set of recurrence relations for the calculation of the probabilities of a large class of integer-valued random variables. We show that the probability mass function can be recursively computed for random variables with a probability generating function satisfying certain functional form.

Nonconventional limits of random sequences related to partitions of integers

We deal with a sequence of integer-valued random variables $\{Z_N\}_{N=1}^{\infty}$ which is related to restricted partitions of positive integers. We observe that $Z_N=X_1+ \ldots + X_N$ for independent and bounded random variables $X_j$'s, so $Z_N$ has finite mean ${\bf E}Z_N$ and variance ${\bf Var}Z_N$. We want to find the limit distribution of ${\hat Z}_N=\left(Z_N-{\bf E}Z_N\right)/{\sqrt{{\bf Var}Z_N}}$ as $N \to \infty.$ While in many cases the limit distribution is normal, the main results established in this paper are that ${\hat Z}_N \overset{d}{\to} Z_{*},$ where $Z_{*}$ is a bounded random variable. We find explicitly the range of values of $Z_*$ and derive some properties of its distribution. The main tools used are moment generating functions, cumulant generating functions, moments and cumulants of the random variables involved. Useful related topics are also discussed.

The roots of generating functions and sums of integer-valued random variables

Рассматриваются свойства корней производящих функций целочисленных ограниченных случайных величин и свойства сумм независимых случайных величин со значениями в множествах $\{0, 1\}$ и $\{0, 1, 2\}$. Указаны условия сходимости распределений целочисленных ограниченных случайных величин к пуассоновскому и нормальному распределениям в терминах корней производящих функций.

On the accuracy of Poisson approximation

The problem of evaluating the accuracy of Poisson approximation to the distribution of a sum of independent integer-valued random variables has attracted a lot of attention in the past six decades. From a practical point of view, it has important applications in insurance, reliability theory, extreme value theory, etc.; from a theoretical point of view, the topic provides insights into Kolmogorov’s problem. The task of establishing an estimate with the best possible constant at the leading term remained open for decades. The paper presents a solution to that problem. A first-order asymptotic expansion is established as well. We generalise and sharpen the corresponding inequalities of Prokhorov, LeCam, Barbour, Hall, Deheuvels, Pfeifer, and Roos. A new result is established for the intensively studied topic of Poisson approximation to the binomial distribution.

Renewal processes, population dynamics, and unimodular trees

AbstractBased on a simple object, an i.i.d. sequence of positive integer-valued random variables {an}n∊ℤ, we introduce and study two random structures and their connections. First, a population dynamics, in which each individual is born at time n and dies at time n + an. This dynamics is that of a D/GI/∞ queue, with arrivals at integer times and service times given by {an}n∊ℤ. Second, the directed random graph Tf on ℤ generated by the random map f(n) = n + an. Assuming only that E [a0] &lt; ∞ and P [a0 = 1] &gt; 0, we show that, in steady state, the population dynamics is regenerative, with one individual alive at each regeneration epoch. We identify a unimodular structure in this dynamics. More precisely, Tf is a unimodular directed tree, in which f(n) is the parent of n. This tree has a unique bi-infinite path. Moreover, Tf splits the integers into two categories: ephemeral integers, with a finite number of descendants of all degrees, and successful integers, with an infinite number. Each regeneration epoch is a successful individual such that all integers less than it are its descendants of some order. Ephemeral, successful, and regeneration integers form stationary and mixing point processes on ℤ.

Majorization and Rényi entropy inequalities via Sperner theory

A natural link between the notions of majorization and strongly Sperner posets is elucidated. It is then used to obtain a variety of consequences, including new Rényi entropy inequalities for sums of independent, integer-valued random variables.