Comment on ‘An improved score interval with a modified midpoint for a binomial proportion’

Yu et al. [An improved score interval with a modified midpoint for a binomial proportion. J Stat Comput Simul. 2014;84:1022–1038] propose a novel confidence interval (CI) for a binomial proportion by modifying the midpoint of the score interval. This CI is competitive with the various commonly used methods. At the same time, Martín and Álvarez [Two-tailed asymptotic inferences for a proportion. J Appl Stat. 2014;41:1516–1529] analyse the performance of 29 asymptotic two-tailed CI for a proportion. The CI they selected is based on the arcsin transformation (when this is applied to the data increased by 0.5), although they also refer to the good behaviour of the classical methods of score and Agresti and Coull (which may be preferred in certain circumstances). The aim of this commentary is to compare the four methods referred to previously. The conclusion (for the classic error α of 5%) is that with a small sample size (≤80) the method that should be used is that of Yu et al.; for a large sample size (n?≥?100), the four methods perform in a similar way, with a slight advantage for the Agresti and Coull method. In any case the Agresti and Coull method does not perform badly and tends to be conservative. The program which determines these four intervals are available from the address http://www.ugr.es/local/bioest/Z_LINEAR_K.EXEhttp://www.ugr.es/local/bioest/Z_LINEAR_K.EXE.     

On Some New α-Modified Binomial and Poisson Distributions and Their Applications

A new class of α-modified binomial distribution has been proposed, and its distributional properties like probability generating function (pgf), moments, and their interrelations have been studied. Two new α-modified Poisson distributions and Poisson distribution have been obtained as limiting distributions. Modified binomial and Poisson distributions introduced by Berg and Jaworski (1988 Berg , S. , Jaworski , J. ( 1988 ). Modified binomial and Poisson distributions with application in random mapping theory . J. Statist. Plann. Infer. 18 : 313 – 322 . [Google Scholar]) have been seen as particular cases. Mixture distributions of α-modified binomial distributions have been derived. A new distributions called α-modified binomial distributions of type j, their moment properties, limiting distributions as α-modified Poisson distribution of type j, their different convolution properties, pgf, parameter estimators have been studied. Two more new distributions namely Doubly α-modified binomial distributions of type (i, j) and α-modified weighted generalized Poisson distributions of type (j ? 1) have also been studied. Various α-modified binomial and Poisson distributions of Berg and Mutafchiev (1990 Berg , S. , Mutafchiev , L. ( 1990 ). Random mapping with an attracting center: Lagrangian distributions and a regression function . J. Appl. Probab. 27 : 622 – 636 . [Google Scholar]) and Berg and Nowicki (1991 Berg , S. , Nowicki , K. ( 1991 ). Statistical inference for a class of modified power series distributions with applications to random mapping theory . J. Statist. Plann. Infer. 28 : 247 – 261 . [Google Scholar]) have been seen as special cases. Application of some of these proposed distributions have been identified.     

An approximation to the convolution of gamma distributions

In general, the exact distribution of a convolution of independent gamma random variables is quite complicated and does not admit a closed form. Of all the distributions proposed, the gamma-series representation of Moschopoulos (1985 Moschopoulos, P. G. (1985). The distribution of the sum of independent gamma random variables. Annals of the Institute of Statistical Mathematics 37Part A:541–544. [Google Scholar]) is relatively simple to implement but for particular combinations of scale and/or shape parameters the computation of the weights of the series can result in complications with too much time consuming to allow a large-scale application. Recently, a compact random parameter representation of the convolution has been proposed by Vellaisamy and Upadhye (2009 Vellaisamy, P., Upadhye, N. S. (2009). On the sums of compound negative binomial and gamma random variables. Journal of Applied Probability 46:272–283.[Crossref], [Web of Science ®] , [Google Scholar]) and it allows to give an exact interpretation to the weights of the series. They describe an infinite discrete probability distribution. This result suggested to approximate Moschopoulos’s expression looking for an approximating theoretical discrete distribution for the weights of the series. More precisely, we propose a general negative binomial distribution. The result is an “excellent” approximation, fast and simple to implement for any parameter combination.     

On the performance of the poisson and the negative binomial ridge predictors

Mansson and Shukur (2011 Mansson, K., Shukur, G. (2011). A Poisson ridge regression estimator. Economic Modelling 28:1475–1481. [Google Scholar]) investigated the performance of the Poisson ridge regression (PRR) estimator in terms of the mean square error (MSE) criterion. Similarly, Mansson (2012 Mansson, K. (2012). On ridge estimators for the negative binomial regression model. Economic Modelling 29:178–184. [Google Scholar]) investigated the performance of the Negative binomial ridge regression (NBRR) according to the MSE criterion. But there is no any analysis of the predictive performance of the PRR and NBRR estimators. Therefore, we define the PRR and the NBRR predictors to evaluate their predictive performances according to the prediction mean squared error under the target function. The Monte Carlo simulations and the real life numerical example are conducted to investigate the defined predictors' performance.     

Correlation Structure and Model Selection for Negative Binomial Distribution in GEE

A new model selection criterion, termed as the “quasi-likelihood under the independence model criterion” (QIC), was proposed by Pan (2001 Pan , W. ( 2001 ). Akaike's information criterion in generalized estimating equations . Biometrics 57 : 120 – 125 .[Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) for GEE models. Cui (2007 Cui , J. ( 2007 ). QIC program and model selection in GEE analyses . Stata Journal 7 : 209 – 220 .[Web of Science ®] , [Google Scholar]) developed a general computing program to implement the QIC method for a range of statistical distributions. However, only a special case of the negative binomial distribution was considered in Cui (2007 Cui , J. ( 2007 ). QIC program and model selection in GEE analyses . Stata Journal 7 : 209 – 220 .[Web of Science ®] , [Google Scholar]), where the dispersion parameter equals to unity. This article introduces a new computing program that can be applied for the general negative binomial model, where the dispersion parameter can be any fixed value. An example is also given in this article.     

Optimal sample size for record data and associated cost analysis for exponential distribution

Estimation of the mean of an exponential distribution based on record data has been treated by Samaniego and Whitaker [F.J. Samaniego, and L.R. Whitaker, On estimating popular characteristics from record breaking observations I. Parametric results, Naval Res. Logist. Quart. 33 (1986), pp. 531–543] and Doostparast [M. Doostparast, A note on estimation based on record data, Metrika 69 (2009), pp. 69–80]. When a random sample Y ₁, …, Y _n is examined sequentially and successive minimum values are recorded, Samaniego and Whitaker [F.J. Samaniego, and L.R. Whitaker, On estimating popular characteristics from record breaking observations I. Parametric results, Naval Res. Logist. Quart. 33 (1986), pp. 531–543] obtained a maximum likelihood estimator of the mean of the population and showed its convergence in probability. We establish here its convergence in mean square error, which is stronger than the convergence in probability. Next, we discuss the optimal sample size for estimating the mean based on a criterion involving a cost function as well as the Fisher information based on records arising from a random sample. Finally, a comparison between complete data and record is carried out and some special cases are discussed in detail.     

Stirling's Formula and Its Extensions: Heuristic Approaches

Walsh (1995 Walsh , D. P. ( 1995 ). Equating Poisson and normal probability functions to derive Stirling's formula . Amer. Statist. 49 : 270 – 271 .[Taylor & Francis Online] , [Google Scholar]) introduced a heuristic approach to motivate Stirling's formula by equating a Poisson probability to an analogous value from a normal density function. We explore similar heuristics to derive approximations for various binomial, negative binomial, and multinomial coefficients. Also, using heuristics markedly different from those of Walsh, we develop an approximation of (nk)! for positive integers n (large) and k. These heuristics are then used to validate Stirling's formula for Γ(nα) where α is a positive real number. To derive each of our approximations we use a different probability distribution, and hence each section may serve as pedagogical module.     

A Bayesian Inference and Stochastic Dynamic Programming Approach to Determine the Best Binomial Distribution

In this research, we employ Bayesian inference and stochastic dynamic programming approaches to select the binomial population with the largest probability of success from n independent Bernoulli populations based upon the sample information. To do this, we first define a probability measure called belief for the event of selecting the best population. Second, we explain the way to model the selection problem using Bayesian inference. Third, we clarify the model by which we improve the beliefs and prove that it converges to select the best population. In this iterative approach, we update the beliefs by taking new observations on the populations under study. This is performed using Bayesian rule and prior beliefs. Fourth, we model the problem of making the decision in a predetermined number of decision stages using the stochastic dynamic programming approach. Finally, in order to understand and to evaluate the proposed methodology, we provide two numerical examples and a comparison study by simulation. The results of the comparison study show that the proposed method performs better than that of Levin and Robbins (1981 Levin , B. , Robbins , H. ( 1981 ). Selecting the highest probability in Binomial or multinomial trials . Proc. Nat. Acad. Sci. USA 78 : 4663 – 4666 .[Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) for some values of estimated probability of making a correct selection.     

New large-sample confidence intervals for a linear combination of binomial proportions

In this paper, we consider the problem wherein one desires to estimate a linear combination of binomial probabilities from k>2$k>2$ independent populations. In particular, we create a new family of asymptotic confidence intervals, extending the approach taken by Beal [1987. Asymptotic confidence intervals for the difference between two binomial parameters for use with small samples. Biometrics 73, 941–950] in the two-sample case. One of our new intervals is shown to perform very well when compared to the best available intervals documented in Price and Bonett [2004. An improved confidence interval for a linear function of binomial proportions. Comput. Statist. Data Anal. 45, 449–456]. Furthermore, our interval estimation approach is quite general and could be extended to handle more complicated parametric functions and even to other discrete probability models in stratified settings. We illustrate our new intervals using two real data examples, one from an ecology study and one from a multicenter clinical trial.     

Bayesian Analysis of Poisson Regression with Lognormal Unobserved Heterogeneity: With an Application to the Patent-R&D Relationship

This article considers explicit and detailed theoretical and empirical Bayesian analysis of the well-known Poisson regression model for count data with unobserved individual effects based on the lognormal, rather than the popular negative binomial distribution. Although the negative binomial distribution leads to analytical expressions for the likelihood function, a Poisson-lognormal model is closer to the concept of regression with normally distributed innovations, and accounts for excess zeros as well. Such models have been considered widely in the literature (Winkelmann, 2008 Winkelmann , R. ( 2008 ). Econometric Analysis of Count Data. , 5th ed. Berlin : Springer . [Google Scholar]). The article also provides the necessary theoretical results regarding the posterior distribution of the model. Given that the likelihood function involves integrals with respect to the latent variables, numerical methods organized around Gibbs sampling with data augmentation are proposed for likelihood analysis of the model. The methods are applied to the patent-R&D relationship of 70 US pharmaceutical and biomedical companies, and it is found that it performs better than Poisson regression or negative binomial regression models.     

Non-parametric prediction intervals for the lifetime of coherent systems

In this paper, nonparametric methods are proposed to construct prediction intervals for the lifetime of a coherent system with known signatures. An explicit expression for the coverage probability of the prediction intervals is presented based on Samaniego’s signature. The existence and optimality of these intervals are discussed. In our derivation, we also obtain an exact expression for the marginal distribution of the \(i\) th order statistic from a pooled sample.     

On the beta-correlated binomial (bcb) distribution - a three parameter generalization of the binomial distribution

A modification of the beta-correlated binomial (BCB) distribution of Paul (1985) is proposed. This modification over- comes some theoretical difficulties encountered in the original BCB distribution.     

Confidence interval estimation for negative binomial group distribution

Negative binomial group distribution was proposed in the literature which was motivated by inverse sampling when considering group inspection: products are inspected group by group, and the number of non-conforming items of a group is recorded only until the inspection of the whole group is finished. The non-conforming probability p of the population is thus the parameter of interest. In this paper, the confidence interval construction for this parameter is investigated. The common normal approximation and exact method are applied. To overcome the drawbacks of these commonly used methods, a composite method that is based on the confidence intervals of the negative binomial distribution is proposed, which benefits from the relationship between negative binomial distribution and negative binomial group distribution. Simulation studies are carried out to examine the performances of our methods. A real data example is also presented to illustrate the application of our method.     

A backward construction and simulation of correlated Poisson processes

In this paper, we consider a generalisation of the backward simulation method of Duch et al. [New approaches to operational risk modeling. IBM J Res Develop. 2014;58:1–9] to build bivariate Poisson processes with flexible time correlation structures, and to simulate the arrival times of the processes. The proposed backward construction approach uses the Marshall–Olkin bivariate binomial distribution for the conditional law and some well-known families of bivariate copulas for the joint success probability in lieu of the typical conditional independence assumption. The resulting bivariate Poisson process can exhibit various time correlation structures which are commonly observed in real data.     

Visualizing Count Data Regressions Using Rootograms

The rootogram is a graphical tool associated with the work of J. W. Tukey that was originally used for assessing goodness of fit of univariate distributions. Here, we extend the rootogram to regression models and show that this is particularly useful for diagnosing and treating issues such as overdispersion and/or excess zeros in count data models. We also introduce a weighted version of the rootogram that can be applied out of sample or to (weighted) subsets of the data, for example, in finite mixture models. An empirical illustration revisiting a well-known dataset from ethology is included, for which a negative binomial hurdle model is employed. Supplementary materials providing two further illustrations are available online: the first, using data from public health, employs a two-component finite mixture of negative binomial models; the second, using data from finance, involves underdispersion. An R implementation of our tools is available in the R package countreg. It also contains the data and replication code.     

Analysis of hypoglycemic events using negative binomial models

Negative binomial regression is a standard model to analyze hypoglycemic events in diabetes clinical trials. Adjusting for baseline covariates could potentially increase the estimation efficiency of negative binomial regression. However, adjusting for covariates raises concerns about model misspecification, in which the negative binomial regression is not robust because of its requirement for strong model assumptions. In some literature, it was suggested to correct the standard error of the maximum likelihood estimator through introducing overdispersion, which can be estimated by the Deviance or Pearson Chi‐square. We proposed to conduct the negative binomial regression using Sandwich estimation to calculate the covariance matrix of the parameter estimates together with Pearson overdispersion correction (denoted by NBSP). In this research, we compared several commonly used negative binomial model options with our proposed NBSP. Simulations and real data analyses showed that NBSP is the most robust to model misspecification, and the estimation efficiency will be improved by adjusting for baseline hypoglycemia. Copyright © 2013 John Wiley & Sons, Ltd.