Lower-Bound Estimates for Poisson-Type Approximations

Let S = w ₁ S ₁ + w ₂ S ₂ + ⋯ + w _N S _N . Here S _j is a sum of identically distributed random variables with weight w _j > 0. We consider the cases where S _j is a sum of independent random variables, the sum of independent lattice variables, or has the Markov binomial distribution. Apart from the general case, we investigate the case of symmetric random variables. Distribution of S is approximated by a compound Poisson distribution, by a second-order asymptotic expansion, and by a signed exponential measure. Lower bounds for the accuracy of approximations in uniform metric are established. __________ Translated from Lietuvos Matematikos Rinkinys, Vol. 45, No. 4, pp. 501–524, October–December, 2005.     

Exact Simulation of IG-OU Processes

IG-OU processes are a subclass of the non-Gaussian processes of Ornstein–Uhlenbeck type, which are important models appearing in financial mathematics and elsewhere. The simulation of these processes is of interest for its applications in statistical inference. In this paper, a stochastic integral of Ornstein–Uhlenbeck type is represented to be the sum of two independent random variables—one has an inverse Gaussian distribution and the other has a compound Poisson distribution. And in distribution, the compound Poisson random variable is equal to a sum of Poisson-distributed number positive random variables, which are independent identically distributed and have a common specified density function. The exact simulation of the IG-OU processes, proceeding from time 0 and going in steps of time interval Δ, is achieved via the representation of the stochastic integral. Comparing to the approximate method, which is based on Rosinski’s infinite series representation of the same stochastic integral, by the quantile–quantile plots, the advantage of the exact simulation method is obvious. In addition, as an application, we provide an estimator of the intensity parameter of the IG-OU processes and validate its superiority to another estimator by our exact simulation method.      

Nonparametric estimation of compound distributions with applications in insurance

A nonparametric estimator of the distribution functionG of a random sum of independent identically distributed random variables, with distribution functionF, is proposed in the case where the distribution of the number of summands is known and a random sample fromF is available. This estimator is found by evaluating the functional that mapsF ontoG at the empirical distribution function based on the random sample. Strong consistency and asymptotic normality of the resulting estimator in a suitable function space are established using appropriate continuity and differentiability results for the functional. Bootstrap confidence bands are also obtained. Applications to the aggregate claims distribution function and to the probability of ruin in the Poisson risk model are presented.     

Inference on a distribution function from ranked set samples

Consider independent observations \((X_i,R_i)\) with random or fixed ranks \(R_i\), while conditional on \(R_i\), the random variable \(X_i\) has the same distribution as the \(R_i\)-th order statistic within a random sample of size k from an unknown distribution function F. Such observation schemes are well known from ranked set sampling and judgment post-stratification. Within a general, not necessarily balanced setting we derive and compare the asymptotic distributions of three different estimators of the distribution function F: a stratified estimator, a nonparametric maximum-likelihood estimator and a moment-based estimator. Our functional central limit theorems generalize and refine previous asymptotic analyses. In addition, we discuss briefly pointwise and simultaneous confidence intervals for the distribution function with guaranteed coverage probability for finite sample sizes. The methods are illustrated with a real data example, and the potential impact of imperfect rankings is investigated in a small simulation experiment.

     

Poisson and compound Poisson distributions of order k and some of their properties

One investigates the Poisson distribution of orderk. One finds its generating function and with its aid one establishes that the sum of independent random variables, distributed according to the Poisson law of orderk, is distributed in the same manner. In addition, one considers a generalized compound Poisson distribution.Translated from Zapiski Nauchnykh Seminarov Leningradskogo Otdeleniya Matematicheskogo Instituta im. V. A. Steklova AN SSSR, Vol. 130, pp. 175–180, 1983.     

Some properties of a two-parameter family of poisson distributions of orderk

We study the problem of simulating the process of detecting a weighted sum of independent Poisson random variables. We investigate the properties of the resulting compound Poisson distribution: the analytic form of the probability function and recursion formulas for computing it, moments and semi-invariants, asymptotics of the distribution, and recursion relations for the derivatives with respect to the parameters. We give the results of model computations showing the set structure of the distribution. One figure. Bibliography: 8 titles. Translated fromProblemy Matematicheskoi Fiziki, 1998, pp. 46–54.     

Gambler’s ruin and winning a series by <Emphasis Type="Italic">m</Emphasis> games

Two teams play a series of games until one team accumulates m more wins than the other. These series are fairly common in some sports provided that the competition has already extended beyond some number of games. We generalize these schemes to allow ties in the single games. Different approaches offer different advantages in calculating the winning probabilities and the distribution of the duration N, including difference equations, conditioning, explicit and implicit path counting, generating functions and a martingale-based derivation of the probability and moment generating functions of N. The main result of the paper is the determination of the exact distribution of N for a series of fair games without ties as a sum of independent geometrically distributed random variables and its approximation.     

Poisson approximation for large deviations

Upper and lower bounds are given for P(S ≤ k), 0 ≤ k ≤ ES, where S is a sum of indicator variables with a special structure, which appears, for example, in subgraph counts in random graphs. in typical cases, these bounds are close to the corresponding probabilities for a Poisson distribution with the same mean as S. There are no corresponding general bounds for P(S ≥ k), k > ES, but some partial results are given.     

Nonparametric Estimation of Regression Functions in Point Process Models

We prove that the empirical L ²-risk minimizing estimator over some general type of sieve classes is universally, strongly consistent for the regression function in a class of point process models of Poissonian type (random sampling processes). The universal consistency result needs weak assumptions on the underlying distributions and regression functions. It applies in particular to neural net classes and to radial basis function nets. For the estimation of the intensity functions of a Poisson process a similar technique yields consistency of the sieved maximum likelihood estimator for some general sieve classes. This revised version was published online in August 2006 with corrections to the Cover Date.     

Poisson convergence in the restricted k‐partitioning problem

The randomized k‐number partitioning problem is the task to distribute N i.i.d. random variables into k groups in such a way that the sums of the variables in each group are as similar as possible. The restricted k‐partitioning problem refers to the case where the number of elements in each group is fixed to N/k. In the case k = 2 it has been shown that the properly rescaled differences of the two sums in the close to optimal partitions converge to a Poisson point process, as if they were independent random variables. We generalize this result to the case k > 2 in the restricted problem and show that the vector of differences between the k sums converges to a k ‐ 1‐dimensional Poisson point process. © 2006 Wiley Periodicals, Inc. Random Struct. Alg., 2007     

Admissible and minimax multiparameter estimation in exponential families

Consider p independent distributions each belonging to the one parameter exponential family with distribution functions absolutely continuous with respect to Lebesgue measure. For estimating the natural parameter vector with p ≥ p₀ (p₀ is typically 2 or 3), a general class of estimators dominating the minimum variance unbiased estimator (MVUE) or an estimator which is a known constant multiple of the MVUE is produced under different weighted squared error losses. Included as special cases are some results of Hudson [13] and Berger [5]. Also, for a subfamily of the general exponential family, a class of estimators dominating the MVUE of the mean vector or an estimator which is a known constant multiple of the MVUE is produced. The major tool is to obtain a general solution to a basic differential inequality.     

An asymptotic expansion for the normalizing constant of the Conway–Maxwell–Poisson distribution

The Conway–Maxwell–Poisson distribution is a two-parameter generalization of the Poisson distribution that can be used to model data that are under- or over-dispersed relative to the Poisson distribution. The normalizing constant \(Z(\lambda ,\nu )\) is given by an infinite series that in general has no closed form, although several papers have derived approximations for this sum. In this work, we start by using probabilistic argument to obtain the leading term in the asymptotic expansion of \(Z(\lambda ,\nu )\) in the limit \(\lambda \rightarrow \infty \) that holds for all \(\nu >0\). We then use an integral representation to obtain the entire asymptotic series and give explicit formulas for the first eight coefficients. We apply this asymptotic series to obtain approximations for the mean, variance, cumulants, skewness, excess kurtosis and raw moments of CMP random variables. Numerical results confirm that these correction terms yield more accurate estimates than those obtained using just the leading-order term.

     

On stable estimation of a function of a parameter

Given the function f and the vector-statistic t_N which is a mean square consistent estimator of a parameter a, the problem is to estimate f(a). The criteria for the mean square consistency of the estimator f(t_N) are considered. In the case where the estimator f(t_N) is not mean square consistent, a class of estimators of f(a) is proposed, and it is proved that the estimators of the class are mean square consistent for all distribution of t_N. Translated fromStatisticheskie Metody Otsenivaniya i Proverki Gipotez, pp. 44–55, Perm, 1990.     

The Arcsine Law

Let N _n denote the number of positive sums in the first n trials in a random walk (S _i) and let L _n denote the first time we obtain the maximum in S ₀,..., S _n. Then the classical equivalence principle states that N _n and L _n have the same distribution and the classical arcsine law gives necessary and sufficient condition for (1/n) L _n or (1/n) N _n to converge in law to the arcsine distribution. The objective of this note is to provide a simple and elementary proof of the arcsine law for a general class of integer valued random variables (T _n) and to provide a simple an elementary proof of the equivalence principle for a general class of integer valued random vectors (N _n, L _n).     

Estimator selection with respect to Hellinger-type risks

We observe a random measure N and aim at estimating its intensity s. This statistical framework allows to deal simultaneously with the problems of estimating a density, the marginals of a multivariate distribution, the mean of a random vector with nonnegative components and the intensity of a Poisson process. Our estimation strategy is based on estimator selection. Given a family of estimators of s based on the observation of N, we propose a selection rule, based on N as well, in view of selecting among these. Little assumption is made on the collection of estimators and their dependency with respect to the observation N need not be known. The procedure offers the possibility to deal with various problems among which model selection, convex aggregation and construction of T-estimators as studied recently in Birgé (Ann Inst H Poincaré Probab Stat 42(3):273?C325, 2006). For illustration, we shall consider the problems of estimation, complete variable selection and selection among linear estimators in possibly non-Gaussian regression settings.     

Extreme Value Distributions for Random Coupon Collector and Birthday Problems

Take n independent copies of a strictly positive random variable X and divide each copy with the sum of the copies, thus obtaining n random probabilities summing to one. These probabilities are used in independent multinomial trials with n outcomes. Let N _n(N ^* _n) be the number of trials needed until each (some) outcome has occurred at least c times. By embedding the sampling procedure in a Poisson point process the distributions of N _n and N ^* _n can be expressed using extremes of independent identically distributed random variables. Using this, asymptotic distributions as n are obtained from classical extreme value theory. The limits are determined by the behavior of the Laplace transform of X close to the origin or at infinity. Some examples are studied in detail.     

COM-negative binomial distribution: modeling overdispersion and ultrahigh zero-inflated count data

We focus on the COM-type negative binomial distribution with three parameters, which belongs to COM-type (a, b, 0) class distributions and family of equilibrium distributions of arbitrary birth-death process. Besides, we show abundant distributional properties such as overdispersion and underdispersion, log-concavity, log-convexity (infinite divisibility), pseudo compound Poisson, stochastic ordering, and asymptotic approximation. Some characterizations including sum of equicorrelated geometrically distributed random variables, conditional distribution, limit distribution of COM-negative hypergeometric distribution, and Stein’s identity are given for theoretical properties. COM-negative binomial distribution was applied to overdispersion and ultrahigh zero-inflated data sets. With the aid of ratio regression, we employ maximum likelihood method to estimate the parameters and the goodness-of-fit are evaluated by the discrete Kolmogorov-Smirnov test.     

Complex random energy model: zeros and fluctuations

The partition function of the random energy model at inverse temperature $\beta $ is a sum of random exponentials $ \mathcal{Z }_N(\beta )=\sum _{k=1}^N \exp (\beta \sqrt{n} X_k)$ , where $X_1,X_2,\ldots $ are independent real standard normal random variables (=random energies), and $n=\log N$ . We study the large N limit of the partition function viewed as an analytic function of the complex variable $\beta $ . We identify the asymptotic structure of complex zeros of the partition function confirming and extending predictions made in the theoretical physics literature. We prove limit theorems for the random partition function at complex $\beta $ , both on the logarithmic scale and on the level of limiting distributions. Our results cover also the case of the sums of independent identically distributed random exponentials with any given correlations between the real and imaginary parts of the random exponent.     

Optimal Upper and Lower Bounds for the Upper Tails of Compound Poisson Processes

A compound Poisson process is of the form where Z, Z ₁, Z ₂, are arbitrary i.i.d. random variables and N is an independent Poisson random variable with parameter . This paper identifies the degree of precision that can be achieved when using exponential bounds together with a single truncation to approximate . The truncation level introduced depends only on and Z and not on the overall exceedance level a.     

Perishable Inventory System at Service Facilities with N Policy

Abstract

This article presents a perishable stochastic inventory system under continuous review at a service facility in which the waiting hall for customers is of finite size M. The service starts only when the customer level reaches N (< M), once the server has become idle for want of customers. The maximum storage capacity is fixed as S. It is assumed that demand for the commodity is of unit size. The arrivals of customers to the service station form a Poisson process with parameter λ. The individual customer is issued a demanded item after a random service time, which is distributed as negative exponential. The items of inventory have exponential life times. It is also assumed that lead time for the reorders is distributed as exponential and is independent of the service time distribution. The demands that occur during stock out periods are lost.The joint probability distribution of the number of customers in the system and the inventory levels is obtained in steady state case. Some measures of system performance in the steady state are derived. The results are illustrated with numerical examples.