Showing papers in "Methodology and Computing in Applied Probability in 2007"

Statistical Process Control using Shewhart Control Charts with Supplementary Runs Rules

Abstract: The aim of this paper is to present the basic principles and recent advances in the area of statistical process control charting with the aid of runs rules. More specifically, we review the well known Shewhart type control charts supplemented with additional rules based on the theory of runs and scans. The motivation for this article stems from the fact that during the last decades, the performance improvement of the Shewhart charts by exploiting runs rules has attracted continuous research interest. Furthermore, we briefly discuss the Markov chain approach which is the most popular technique for studying the run length distribution of run based control charts.

A Survey of the Coupon Collector’s Problem with Random Sample Sizes

Abstract: This paper surveys the coupon collector’s waiting time problem with random sample sizes and equally likely balls. Consider an urn containing m red balls. For each draw, a random number of balls are removed from the urn. The group of removed balls is painted white and returned to the urn. Several approaches to addressing this problem are discussed, including a Markov chain approach to compute the distribution and expected value of the number of draws required for the urn to contain j white balls given that it currently contains i white balls. As a special case, E[N], the expected number of draws until all the balls are white given that all are currently red is also obtained.

An Algorithmic Approach to Discrete Time Non-homogeneous Backward Semi-Markov Reward Processes with an Application to Disability Insurance

Abstract: In this paper semi-Markov reward models are presented. Higher moments of the reward process is presented for the first time applied to in time non-homogeneous semi-Markov insurance problems. Also an example is presented based on real disability data. Different algorithmic approaches to solve the problem is described.

Parametric and Non Homogeneous Semi-Markov Process for HIV Control

Abstract: In AIDS control, physicians have a growing need to use pragmatically useful and interpretable tools in their daily medical taking care of patients. Semi-Markov process seems to be well adapted to model the evolution of HIV-1 infected patients. In this study, we introduce and define a non homogeneous semi-Markov (NHSM) model in continuous time. Then the problem of finding the equations that describe the biological evolution of patient is studied and the interval transition probabilities are computed. A parametric approach is used and the maximum likelihood estimators of the process are given. A Monte Carlo algorithm is presented for realizing non homogeneous semi-Markov trajectories. As results, interval transition probabilities are computed for distinct times and follow-up has an impact on the evolution of patients.

Crossing Probabilities for Diffusion Processes with Piecewise Continuous Boundaries

Abstract: We propose an approach to compute the boundary crossing probabilities for a class of diffusion processes which can be expressed as piecewise monotone (not necessarily one-to-one) functionals of a standard Brownian motion. This class includes many interesting processes in real applications, e.g., Ornstein-Uhlenbeck, growth processes and geometric Brownian motion with time dependent drift. This method applies to both one-sided and two-sided general nonlinear boundaries, which may be discontinuous. Using this approach explicit formulas for boundary crossing probabilities for certain nonlinear boundaries are obtained, which are useful in evaluation and comparison of various computational algorithms. Moreover, numerical computation can be easily done by Monte Carlo integration and the approximation errors for general boundaries are automatically calculated. Some numerical examples are presented.

Asymptotics of an Efficient Monte Carlo Estimation for the Transition Density of Diffusion Processes

Abstract: Discretized simulation is widely used to approximate the transition density of discretely observed diffusions. A recently proposed importance sampler, namely modified Brownian bridge, has gained much attention for its high efficiency relative to other samplers. It is unclear for this sampler, however, how to balance the trade-off between the number of imputed values and the number of Monte Carlo simulations under a given computing resource. This paper provides an asymptotically efficient allocation of computing resource to the importance sampling approach with a modified Brownian bridge as importance sampler. The optimal trade-off is established by investigating two types of errors: Euler discretization error and Monte Carlo error. The main results are illustrated with two simulated examples.

Bayesian Hierarchical Space–time Modeling of Earthquake Data

Abstract: Stochastic earthquake models are often based on a marked point process approach as for instance presented in Vere-Jones (Int. J. Forecast., 11:503–538, 1995). This gives a fine resolution both in space and time making it possible to represent each earthquake. However, it is not obvious that this approach is advantageous when aiming at earthquake predictions. In the present paper we take a coarse point of view considering grid cells of 0.5 × 0.5°, or about 50 × 50 km, and time periods of 4 months, which seems suitable for predictions. More specifically, we will discuss different alternatives of a Bayesian hierarchical space–time model in the spirit of Wikle et al. (Environ. Ecol. Stat., 5:117–154, 1998). For each time period the observations are the magnitudes of the largest observed earthquake within each grid cell. As data we apply parts of an earthquake catalogue provided by The Northern California Earthquake Data Center where we limit ourselves to the area 32–37° N, 115–120° W for the time period January 1981 through December 1999 containing the Landers and Hector Mine earthquakes of magnitudes, respectively, 7.3 and 7.1 on the Richter scale. Based on space-time model alternatives one step earthquake predictions for the time periods containing these two events for all grid cells are arrived at. The model alternatives are implemented within an MCMC framework in Matlab. The model alternative that gives the overall best predictions based on a standard loss is claimed to give new knowledge on the spatial and time related dependencies between earthquakes. Also considering a specially designed loss using spatially averages of the 90th percentiles of the predicted values distribution of each cell it is clear that the best model predicts the high risk areas rather well. By using these percentiles we believe that one has a valuable tool for defining high and low risk areas in a region in short term predictions.

Some Reward Paths in Semi-Markov Models with Stochastic Selection of the Transition Probabilities

Abstract: In the present paper, the reward paths in non homogeneous semi-Markov systems in discrete time are examined with stochastic selection of the transition probabilities. The mean entrance probabilities and the mean rewards in the course of time are evaluated. Then the rate of the total reward for the homogeneous case is examined and the mean total reward is evaluated by means of p.g.f’s.

A Probabilistic Framework Towards the Parameterization of Association Rule Interestingness Measures

Abstract: In this paper, we first present an original and synthetic overview of the most commonly used association rule interestingness measures. These measures usually relate the confidence of a rule to an independence reference situation. Yet, some relate it to indetermination, or impose a minimum confidence threshold. We propose a systematic generalization of these measures, taking into account a reference point chosen by an expert in order to appreciate the confidence of a rule. This generalization introduces new connections between measures, and leads to the enhancement of some of them. Finally we propose new parameterized possibilities.

Optimum Burn-in Time for a Bathtub Shaped Failure Distribution

Abstract: An important problem in reliability is to define and estimate the optimal burn-in time. For bathtub shaped failure-rate lifetime distributions, the optimal burn-in time is frequently defined as the point where the corresponding mean residual life function achieves its maximum. For this point, we construct an empirical estimator and develop the corresponding statistical inferential theory. Theoretical results are accompanied with simulation studies and applications to real data. Furthermore, we develop a statistical inferential theory for the difference between the minimum point of the corresponding failure rate function and the aforementioned maximum point of the mean residual life function. The difference measures the length of the time interval after the optimal burn-in time during which the failure rate function continues to decrease and thus the burn-in process can be stopped.

On the Normal Approximation for the Distribution of the Number of Simple or Compound Patterns in a Random Sequence of Multi-state Trials

Abstract: Distributions of numbers of runs and patterns in a sequence of multi-state trials have been successfully used in various areas of statistics and applied probability. For such distributions, there are many results on Poisson approximations, some results on large deviation approximations, but no general results on normal approximations. In this manuscript, using the finite Markov chain imbedding technique and renewal theory, we show that the number of simple or compound patterns, under overlap or non-overlap counting, in a sequence of multi-state trials follows a normal distribution. Poisson and large deviation approximations are briefly reviewed.

Selecting the Best Population Using a Test for Equality Based on Minimal Wilcoxon Rank-sum Precedence Statistic

Abstract: In this paper, we first give an overview of the precedence-type test procedures. Then we propose a nonparametric test based on early failures for the equality of two life-time distributions against two alternatives concerning the best population. This procedure utilizes the minimal Wilcoxon rank-sum precedence statistic (Ng and Balakrishnan, 2002, 2004) which can determine the difference between populations based on early (100q%) failures. Hence, this procedure can be useful in life-testing experiments in biological as well as industrial settings. After proposing the test procedure, we derive the exact null distribution of the test statistic in the two-sample case with equal or unequal sample sizes. We also present the exact probability of correct selection under the Lehmann alternative. Then, we generalize the test procedure to the k-sample situation. Critical values for some sample sizes are presented. Next, we examine the performance of this test procedure under a location-shift alternative through Monte Carlo simulations. Two examples are presented to illustrate our test procedure with selecting the best population as an objective.

Distributions of Random Partitions and Their Applications

Abstract: Assume that a random sample of size m is selected from a population containing a countable number of classes (subpopulations) of elements (individuals). A partition of the set of sample elements into (unordered) subsets, with each subset containing the elements that belong to same class, induces a random partition of the sample size m, with part sizes {Z 1,Z 2,...,Z N } being positive integer-valued random variables. Alternatively, if N j is the number of different classes that are represented in the sample by j elements, for j=1,2,...,m, then (N 1,N 2,...,N m ) represents the same random partition. The joint and the marginal distributions of (N 1,N 2,...,N m ), as well as the distribution of $N=\sum^m_{j=1}N_{\!j}$ are of particular interest in statistical inference. From the inference point of view, it is desirable that all the information about the population is contained in (N 1,N 2,...,N m ). This requires that no physical, genetical or other kind of significance is attached to the actual labels of the population classes. In the present paper, combinatorial, probabilistic and compound sampling models are reviewed. Also, sampling models with population classes of random weights (proportions), and in particular the Ewens and Pitman sampling models, on which many publications are devoted, are extensively presented.

On New Multivariate Probability Distributions and Stochastic Processes with System Reliability and Maintenance Applications

Abstract: Two distinct methods for construction of some interesting new classes of multivariate probability densities are described and applied. As common results of both procedures three n-variate pdf classes are obtained. These classes are considered as generalizations of the class of univariate Weibullian, gamma, and multivariate normal pdfs. An example of an application of the obtained n-variate pdfs to the problem of modeling the reliability of multicomponent systems with stochastically dependent life-times of their components is given. Obtaining sequences over n = 2, 3, ... of consistent n-variate pdfs, that obey a relatively simple common pattern, for each n, allows us to extend some of the constructions from random vectors to discrete time stochastic processes. Application of one, so obtained, class of highly non-Markovian, but still sufficiently simple, stochastic processes for modeling maintenance of systems with repair, is presented. These models allow us to describe and analyze repaired systems with histories of all past repairs.

Probing Option Prices for Information

Abstract: We present a methodology for extracting information from option prices when the market is viewed as knowledgeable. By expanding the information filtration judiciously and determining conditional characteristic functions for the log of the stock price, we obtain option pricing formulae which when fit to market data may reveal this information. In particular, we consider probing option prices for knowledge of the future stock price, instantaneous volatility, and the asymptotic dividend stream. Additionally the bridge laws developed are also useful for simulation based on stratified sampling that conditions on the terminal values of paths.

Asymptotics for Ratios with Applications to Reinsurance

Abstract: This paper provides an overview of results pertaining to moment convergence for certain ratios of random variables involving sums, order statistics and extreme terms in the sense of modulus. Most of the literature on this matter originates from Darling (1952) who gave a criterion for the convergence in probability to 1 of the ratio of the maximum to the sum in case of nonnegative random variables.

Estimating the Scale Parameter of a Lévy-stable Distribution via the Extreme Value Approach

Abstract: The characteristic exponent α of a Levy-stable law Sα(σ, β, μ) was thor- oughly studied as the extreme value index of a heavy tailed distribution. For 1 <α< 2, Peng (Statist. Probab. Lett. 52: 255-264, 2001) has proposed, via the ex- treme value approach, an asymptotically normal estimator for the location parameter μ. In this paper, we derive by the same approach, an estimator for the scale parameter σ and we discuss its limiting behavior.

Supermodular Comparison of Time-to-Ruin Random Vectors

Abstract: This paper studies time-to-ruin random vectors for multivariate risk processes. Two cases are considered: risk processes with independent increments and risk processes evolving in a common random environment (e.g., because they share the same economic conditions). As expected, increasing the dependence between the risk processes increases the dependence between their respective time-to-ruin random variables.

Stationary Analysis of a Fluid Queue Driven by Some Countable State Space Markov Chain

Abstract: Motivated by queueing systems playing a key role in the performance evaluation of telecommunication networks, we analyze in this paper the stationary behavior of a fluid queue, when the instantaneous input rate is driven by a continuous-time Markov chain with finite or infinite state space. In the case of an infinite state space and for particular classes of Markov chains with a countable state space, such as quasi birth and death processes or Markov chains of the G/M/1 type, we develop an algorithm to compute the stationary probability distribution function of the buffer level in the fluid queue. This algorithm relies on simple recurrence relations satisfied by key characteristics of an auxiliary queueing system with normalized input rates.

Discrete Compound Poisson Process with Curved Boundaries: Polynomial Structures and Recursions

Abstract: This paper provides a review of recent results, most of them published jointly with Ph. Picard, on the exact distribution of the first crossing of a Poisson or discrete compound Poisson process through a given nondecreasing boundary, of curved or linear shape. The key point consists in using an underlying polynomial structure to describe the distribution, the polynomials being of generalized Appell type for an upper boundary and of generalized Abel–Gontcharoff type for a lower boundary. That property allows us to obtain simple and efficient recursions for the numerical determination of the distribution.

A Monte Carlo Estimation of the Entropy for Markov Chains

Abstract: We introduce an estimate of the entropy \(\mathbb{E}_{p^t}(\log p^t)\) of the marginal density p t of a (eventually inhomogeneous) Markov chain at time t≥1. This estimate is based on a double Monte Carlo integration over simulated i.i.d. copies of the Markov chain, whose transition density kernel is supposed to be known. The technique is extended to compute the external entropy \(\mathbb{E}_{p_1^t}(\log p^t)\), where the p 1 t s are the successive marginal densities of another Markov process at time t. We prove, under mild conditions, weak consistency and asymptotic normality of both estimators. The strong consistency is also obtained under stronger assumptions. These estimators can be used to study by simulation the convergence of p t to its stationary distribution. Potential applications for this work are presented: (1) a diagnostic by simulation of the stability property of a Markovian dynamical system with respect to various initial conditions; (2) a study of the rate in the Central Limit Theorem for i.i.d. random variables. Simulated examples are provided as illustration.

Limit Theorems for Sums of Heavy-tailed Variables with Random Dependent Weights

Abstract: Let \(U_{j} ,\;j \in \mathbb{N}\) be independent and identically distributed random variables with heavy-tailed distributions. Consider a sequence of random weights \({\left\{ {W_{j} } \right\}}_{{j \in \mathbb{N}}}\), independent of \({\left\{ {U_{j} } \right\}}_{{j \in \mathbb{N}}}\) and focus on the weighted sums \({\sum olimits_{j = 1}^{{\left[ {nt} \right]}} {W_{j} {\left( {U_{j} - \mu } \right)}} }\), where μ involves a suitable centering. We establish sufficient conditions for these weighted sums to converge to non-trivial limit processes, as n→∞, when appropriately normalized. The convergence holds, for example, if \({\left\{ {W_{j} } \right\}}_{{j \in \mathbb{N}}}\) is strictly stationary, dependent, and W 1 has lighter tails than U 1. In particular, the weights W j s can be strongly dependent. The limit processes are scale mixtures of stable Levy motions. We establish weak convergence in the Skorohod J 1-topology. We also consider multivariate weights and show that they converge weakly in the strong Skorohod M 1-topology. The M 1-topology, while weaker than the J 1-topology, is strong enough for the supremum and infimum functionals to be continuous.

First Exit Times for Ordinary and Compound Poisson Processes with Non-linear Boundaries

Abstract: Distributions of the first-exit times from a region with concave upper boundary are discussed for ordinary and compound Poisson processes. Explicit formulae are developed for the case of ordinary Poisson processes. Recursive formulae are given for the compound Poisson case, where the jumps are positive, having discrete or continuous distributions with finite means. Applications to sequential point estimation and insurance are illustrated.

Skewing Around: Relationships Among Classes of Skewed Distributions

Abstract: The relationships among normal hidden truncation models, closed skew normal families, fundamental skew normal families and extended skew normal families are explored. The models of Arnold and Beaver in (Test 11(1):7–54, 2002) include all of these absolutely continuous models. Slightly more general absolutely continuous models are available with the label of selection models in Arellano-Valle and Genton (Journal of Multivariate Analysis 96:93–116, 2005). The hidden truncation paradigm provides a convenient description of models that subsumes the full spectrum of these skewed models, including singular and absolutely continuous versions.

Prediction and Conditional Simulation of a 2D Lognormal Diffusion Random Field

Abstract: This paper describes techniques for estimation, prediction and conditional simulation of two-parameter lognormal diffusion random fields which are diffusions on each coordinate and satisfy a particular Markov property. The estimates of the drift and diffusion coefficients, which characterize the lognormal diffusion random field under certain conditions, are used for obtaining kriging predictors. The conditional simulations are obtained using the estimates of the drift and diffusion coefficients, kriging prediction and unconditional simulation for the lognormal diffusion random field.

Review of Some Functionals of Compound Poisson Processes and Related Stopping Times

Abstract: The paper reviews recent results of D. Perry, W. Stadje and S. Zacks, on functionals of stopping times and the associated compound Poisson process with lower and upper linear boundaries. In particular, formulae of these functionals are explicitly developed for the total expected discounted cost of discarded service in an M/G/1 queue with restricted accessibility; for the expected total discounted waiting cost in an M/G/1 restricted queue; for the shortage, holding and clearing costs in an inventory system with continuous input; for the risk in sequential estimation and for the transform of the busy period when the upper boundary is random.

Symmetry Studies for Data Analysis

Abstract: This paper describes some of the basic applications of the algebraic theory of canonical decomposition to the analysis of data. The notions of structured data and symmetry studies are discussed and applied to demonstrate their role in well known principles of analysis of variance and their applicability in more general experimental settings.

Stability Criteria for Multi-class Queueing Networks with Re-entrant Lines

Abstract: In this paper we show how the stability criteria for the model proposed in MacPhee and Muller (Queueing Syst 52(3):215–229, 2006) can be applied to queueing networks with re-entrant lines. The model considered has Poisson arrival streams, servers that can be configured in various ways, exponential service times and Markov feedback of completed jobs. The stability criteria are expressed in terms of the mean drifts of the process under the various server configurations. For models with re-entrant lines we impose here a boundary sojourn condition to ensure adequate control of the process when one or more queues are empty. We show with some examples, including the generalised Lu–Kumar network discussed in Nino-Mora and Glazebrook (J Appl Probab 37(3):890–899, 2000), how our results can be applied.

Some Information Theoretic Ideas Useful in Statistical Inference

Abstract: In this paper we discuss four information theoretic ideas and present their implications to statistical inference: (1) Fisher information and divergence generating functions, (2) information optimum unbiased estimators, (3) information content of various statistics, (4) characterizations based on Fisher information.

Performance Analysis of a Fluid Production/Inventory Model with State-dependence

Abstract: We study the long-run average performance of a fluid production/ inventory model which alternates between ON periods and OFF periods. During ON periods of random lengths items are added continuously, at some state-dependent rate, to the inventory. During OFF periods the content decreases (again at some state-dependent rate) back to some basic level. We derive the pertinent reward functionals in closed form. For this analysis the steady-state distributions of the stock level process and its jump counterpart are required. In several examples we use the obtained explicit formulas to maximize the long-run average net revenue numerically.