Inference for functions of parameters in discrete distributions based on fiducial approach: Binomial and Poisson cases

In this article, we propose a simple method of constructing confidence intervals for a function of binomial success probabilities and for a function of Poisson means. The method involves finding an approximate fiducial quantity (FQ) for the parameters of interest. A FQ for a function of several parameters can be obtained by substitution. For the binomial case, the fiducial approach is illustrated for constructing confidence intervals for the relative risk and the ratio of odds. Fiducial inferential procedures are also provided for estimating functions of several Poisson parameters. In particular, fiducial inferential approach is illustrated for interval estimating the ratio of two Poisson means and for a weighted sum of several Poisson means. Simple approximations to the distributions of the FQs are also given for some problems. The merits of the procedures are evaluated by comparing them with those of existing asymptotic methods with respect to coverage probabilities, and in some cases, expected widths. Comparison studies indicate that the fiducial confidence intervals are very satisfactory, and they are comparable or better than some available asymptotic methods. The fiducial method is easy to use and is applicable to find confidence intervals for many commonly used summary indices. Some examples are used to illustrate and compare the results of fiducial approach with those of other available asymptotic methods.     

New types of shrinkage estimators of Poisson means under the normalized squared error loss

In estimating p( ? 2) independent Poisson means, Clevenson and Zidek (1975) have proposed a class of estimators that shrink the unbiased estimator to the origin and dominate the unbiased one under the normalized squared error loss. This class of estimators was subsequently enlarged in several directions. This article deals with the problem and proposes new classes of dominating estimators using prior information pertinently. Dominance is shown by partitioning the sample space into disjoint subsets and averaging the loss difference over each subset. Estimation of several Poisson mean vectors is also discussed. Further, simultaneous estimation of Poisson means under order restriction is treated and estimators which dominate the isotonic regression estimator are proposed for some types of order restrictions.     

A Simulation Study of Some Biasing Parameters for the Ridge Type Estimation of Poisson Regression

This article proposes several estimators for estimating the ridge parameter k based on Poisson ridge regression (RR) model. These estimators have been evaluated by means of Monte Carlo simulations. As performance criteria, we have calculated the mean squared error (MSE), the mean value, and the standard deviation of k. The first criterion is commonly used, while the other two have never been used when analyzing Poisson RR. However, these performance criteria are very informative because, if several estimators have an equal estimated MSE, then those with low average value and standard deviation of k should be preferred. Based on the simulated results, we may recommend some biasing parameters that may be useful for the practitioners in the field of health, social, and physical sciences.     

Nonparametric maximum likelihood estimation of population size based on the counting distribution

Summary.  The paper discusses the estimation of an unknown population size n . Suppose that an identification mechanism can identify n _obs cases. The Horvitz–Thompson estimator of n adjusts this number by the inverse of 1− p ₀, where the latter is the probability of not identifying a case. When repeated counts of identifying the same case are available, we can use the counting distribution for estimating p ₀ to solve the problem. Frequently, the Poisson distribution is used and, more recently, mixtures of Poisson distributions. Maximum likelihood estimation is discussed by means of the EM algorithm. For truncated Poisson mixtures, a nested EM algorithm is suggested and illustrated for several application cases. The algorithmic principles are used to show an inequality, stating that the Horvitz–Thompson estimator of n by using the mixed Poisson model is always at least as large as the estimator by using a homogeneous Poisson model. In turn, if the homogeneous Poisson model is misspecified it will, potentially strongly, underestimate the true population size. Examples from various areas illustrate this finding.     

Connections of the Poisson weight function to overdispersion and underdispersion

In this paper, we establish several connections of the Poisson weight function to overdispersion and underdispersion. Specifically, we establish that the logconvexity (logconcavity) of the mean weight function is a necessary and sufficient condition for overdispersion (underdispersion) when the Poisson weight function does not depend on the original Poisson parameter. We also discuss some properties of the weighted Poisson distributions (WPD). We then introduce a notion of pointwise duality between two WPDs and discuss some associated properties. Next, we present some illustrative examples and provide a discussion on various Poisson weight functions used in practice. Finally, some concluding remarks are made.     

A mixed Poisson model with varying element sizes

The data of Bissell [A.F. Bissell, A negative binomial model with varying element sizes, Biometrika 59 (1972) 435–441] are counts of the number of flaws in rolls of textile fabric of different lengths. Given that the simple Poisson model is rejected due to overdispersion, we propose an analysis of the data by means of a general family of discrete laws, indexed by a parameter that distinguishes several well-known distributions. Our methodology provides some insight into the quality of the production and a relatively straightforward way to evaluate the probability distribution of the aggregate cost of the defects, avoiding the evaluation of convolutions.     

A window-limited generalized likelihood ratio test for monitoring Poisson processes with linear drifts

There is gradually increasing attention devoted to the monitoring of Poisson process due to its wide applications in industry quality control and health-care surveillance. However, most of the study focuses on the case with step shifts in Poisson means. Relatively little attention has been paid to the case with linear drifts in Poisson means. This paper extends the window-limited generalized likelihood ratio (WGLR) test from the monitoring of normal means to Poisson processes, with focus on linear drifts. The comparison results with the adaptive cumulative sum (ACUSUM) charts and the weighted CUSUM (WCUSUM) charts show that the WGLR chart generally provides better detection performance than the other alternative methods in both the zero-state and steady-state cases.     

Zero-inflated Poisson and negative binomial integer-valued GARCH models

Zero inflation means that the proportion of 0's of a model is greater than the proportion of 0's of the corresponding Poisson model, which is a common phenomenon in count data. To model the zero-inflated characteristic of time series of counts, we propose zero-inflated Poisson and negative binomial INGARCH models, which are useful and flexible generalizations of the Poisson and negative binomial INGARCH models, respectively. The stationarity conditions and the autocorrelation function are given. Based on the EM algorithm, the estimating procedure is simple and easy to be implemented. A simulation study shows that the estimation method is accurate and reliable as long as the sample size is reasonably large. A real data example leads to superior performance of the proposed models compared with other competitive models in the literature.     

Improvements in the Poisson approximation of mixed Poisson distributions

We consider the approximation of mixed Poisson distributions by Poisson laws and also by related finite signed measures of higher order. Upper bounds and asymptotic relations are given for several distances. Even in the case of the Poisson approximation with respect to the total variation distance, our bounds have better order than those given in the literature. In particular, our results hold under weaker moment conditions for the mixing random variable. As an example, we consider the approximation of the negative binomial distribution, which enables us to prove the sharpness of a constant in the upper bound of the total variation distance. The main tool is an integral formula for the difference of the counting densities of a Poisson distribution and a related finite signed measure.     

Estimation in the generalized linear empirical bayes model using the extended quasi-likelihood

A generalized linear empirical Bayes model is developed for empirical Bayes analysis of several means in natural exponential families. A unified approach is presented for all natural exponential families with quadratic variance functions (the Normal, Poisson, Binomial, Gamma, and two others.) The hyperparameters are estimated using the extended quasi-likelihood of Nelder and Pregibon (1987), which is easily implemented via the GLIM package. The accuracy of these estimates is developed by asymptotic approximation of the variance. Two data examples are illustrated.     

A Note on Comparing Several Poisson Means

There has been a renewed interest lately in comparing the means (rates) of several Poisson distributions (processes) due to its wide applicability in various fields, especially in life sciences. In this article we first review the recent developments in the literature, and then propose a parametric bootstrap method which performs as good as, if not better than, the existing methods. Results of our comprehensive simulation study have been provided to compare the relevant methods. Finally these methods have been used to four real-life datasets including a most recent one obtained from a clinical trial on the Intrinsa hormone patch developed by Proctor & Gamble.     

Direct simulation for discrete mixture distributions

We demonstrate how to perform direct simulation for discrete mixture models. The approach is based on directly calculating the posterior distribution using a set of recursions which are similar to those of the Forward-Backward algorithm. Our approach is more practicable than existing perfect simulation methods for mixtures. For example, we analyse 1096 observations from a 2 component Poisson mixture, and 240 observations under a 3 component Poisson mixture (with unknown mixture proportions and Poisson means in each case). Simulating samples of 10,000 perfect realisations took about 17 minutes and an hour respectively on a 900 MHz ultraSPARC computer. Our method can also be used to perform perfect simulation from Markov-dependent mixture models. A byproduct of our approach is that the evidence of our assumed models can be calculated, which enables different models to be compared.     

Tukey’s M-estimator of the Poisson parameter with a special focus on small means

We treat robust M-estimators for independent and identically distributed Poisson data. We introduce modified Tukey M-estimators with bias correction and compare them to M-estimators based on the Huber function as well as to weighted likelihood and other estimators by simulation in case of clean data and data with outliers. In particular, we investigate the problem of combining robustness and high efficiencies at small Poisson means caused by the strong asymmetry of such Poisson distributions and propose a further estimator based on adaptive trimming. The advantages of the constructed estimators are illustrated by an application to smoothing count data with a time varying mean and level shifts.     

Some contributions on the multivariate Poisson–Skellam probability distribution

In this article, we introduce a new form of distribution whose components have the Poisson or Skellam marginal distributions. This new specification allows the incorporation of relevant information on the nature of the correlations between every component. In addition, we present some properties of this distribution. Unlike the multivariate Poisson distribution, it can handle variables with positive and negative correlations. It should be noted that we are only interested in modeling covariances of order 2, which means between all pairs of variables. Some simulations are presented to illustrate the estimation methods. Finally, an application of soccer teams data will highlight the relationship between number of points per season and the goal differential by some covariates.     

BAYESIAN ANALYSIS FOR THE SUPERPOSITION OF TWO DEPENDENT NONHOMOGENEOUS POISSON PROCESSES

ABSTRACT

A Bayesian analysis for the superposition of two dependent nonhomogenous Poisson processes is studied by means of a bivariate Poisson distribution. This particular distribution presents a new likelihood function which takes into account the correlation between the two nonhomogenous Poisson processes. A numerical example using Markov Chain Monte Carlo method with data augmentation is considered.     

Bayesian inference for Poisson and multinomial log-linear models

Categorical data frequently arise in applications in the Social Sciences. In such applications, the class of log-linear models, based on either a Poisson or (product) multinomial response distribution, is a flexible model class for inference and prediction. In this paper we consider the Bayesian analysis of both Poisson and multinomial log-linear models. It is often convenient to model multinomial or product multinomial data as observations of independent Poisson variables. For multinomial data, Lindley (1964) [20] showed that this approach leads to valid Bayesian posterior inferences when the prior density for the Poisson cell means factorises in a particular way. We develop this result to provide a general framework for the analysis of multinomial or product multinomial data using a Poisson log-linear model. Valid finite population inferences are also available, which can be particularly important in modelling social data. We then focus particular attention on multivariate normal prior distributions for the log-linear model parameters. Here, an improper prior distribution for certain Poisson model parameters is required for valid multinomial analysis, and we derive conditions under which the resulting posterior distribution is proper. We also consider the construction of prior distributions across models, and for model parameters, when uncertainty exists about the appropriate form of the model. We present classes of Poisson and multinomial models, invariant under certain natural groups of permutations of the cells. We demonstrate that, if prior belief concerning the model parameters is also invariant, as is the case in a ‘reference’ analysis, then the choice of prior distribution is considerably restricted. The analysis of multivariate categorical data in the form of a contingency table is considered in detail. We illustrate the methods with two examples.     

The analysis of ratios and cross-product ratios of poisson variates with application to incidence rates

The incidence of most diseases is low enough that in. large populations the number of new cases may be considered a Poisson variate. This paper explores models and methods for analyzing such data Specific cases are the estimation and testing of ratios and the cross-product ratios, both simple and stratified* We assume the Poisson means are exponential functions of the relevant parameters. The resulting sets of sufficient statistics are partitioned into a test statistic and a vector of statistics related to the nuisance parameters . The methods derived are based on the conditional distribution of the test statistic given the other sufficient statistics. The analyses of stratified cross-product ratios are seen to be analogues of the noncentral distribution associated with theanalysis of the common odds ratio in several 2×2 tables. The various methods are illustrated in numerical examples involving incidence rates of cancer in two metropolitan areas adjusting for both age and sex.     

Regression for doubly inflated multivariate Poisson distributions

Dependent multivariate count data occur in several research studies. These data can be modelled by a multivariate Poisson or Negative binomial distribution constructed using copulas. However, when some of the counts are inflated, that is, the number of observations in some cells are much larger than other cells, then the copula-based multivariate Poisson (or Negative binomial) distribution may not fit well and it is not an appropriate statistical model for the data. There is a need to modify or adjust the multivariate distribution to account for the inflated frequencies. In this article, we consider the situation where the frequencies of two cells are higher compared to the other cells and develop a doubly inflated multivariate Poisson distribution function using multivariate Gaussian copula. We also discuss procedures for regression on covariates for the doubly inflated multivariate count data. For illustrating the proposed methodologies, we present real data containing bivariate count observations with inflations in two cells. Several models and linear predictors with log link functions are considered, and we discuss maximum likelihood estimation to estimate unknown parameters of the models.     

Financial data modeling by Poisson mixture regression

In many financial applications, Poisson mixture regression models are commonly used to analyze heterogeneous count data. When fitting these models, the observed counts are supposed to come from two or more subpopulations and parameter estimation is typically performed by means of maximum likelihood via the Expectation–Maximization algorithm. In this study, we discuss briefly the procedure for fitting Poisson mixture regression models by means of maximum likelihood, the model selection and goodness-of-fit tests. These models are applied to a real data set for credit-scoring purposes. We aim to reveal the impact of demographic and financial variables in creating different groups of clients and to predict the group to which each client belongs, as well as his expected number of defaulted payments. The model's conclusions are very interesting, revealing that the population consists of three groups, contrasting with the traditional good versus bad categorization approach of the credit-scoring systems.     

The probability function of a geometric Poisson distribution

The geometric Poisson (also called Pólya–Aeppli) distribution is a particular case of the compound Poisson distribution. In this study, the explicit probability function of the geometric Poisson distribution is derived and a straightforward proof for this function is given. By means of a proposed algorithm, some numerical examples and an application on traffic accidents are also given to illustrate the usage of the probability function and proposed algorithm.