Estimation of random coefficient regression models

Linear regression models with coefficients across individual units regarded as random samples from some population are studied in this article from a Bayesian viewpoint. A prior distribution of the secondary parameters is derived following the Jeffreys rule. Posterior distribution of the primary and secondary parameters, and the predictive distribution of the future value are then examined. Computations of the parameter estimates are found to be rather straightforward. Data from a performance test on pigs is analysed and discussed. We also discuss the difficulties involved in using a Lindley and Smith (1972) prior in this problem.     

Theoretical considerations of the multivariate von Mises-Fisher distribution

We develop theorems that are analogous to the famous theorems of Lindley & Smith (1972) for the case that the data are from the p-variate von Mises-Fisher distribution. Results similar to those of Lindley & Smith are obtained for the two-stage case. There is, however, a departure from normal theory in the three-stage case, since closed form solutions do not exist when sampling Sfom the von Mises-Fisher distribution. The difference is discussed and elaborated.     

Higher order moments of order statistics from the Lindley distribution and associated inference

ABSTRACT

The Lindley distribution is an important distribution for analysing the stress–strength reliability models and lifetime data. In many ways, the Lindley distribution is a better model than that based on the exponential distribution. Order statistics arise naturally in many of such applications. In this paper, we derive the exact explicit expressions for the single, double (product), triple and quadruple moments of order statistics from the Lindley distribution. Then, we use these moments to obtain the best linear unbiased estimates (BLUEs) of the location and scale parameters based on Type-II right-censored samples. Next, we use these results to determine the mean, variance, and coefficients of skewness and kurtosis of some certain linear functions of order statistics to develop Edgeworth approximate confidence intervals of the location and scale Lindley parameters. In addition, we carry out some numerical illustrations through Monte Carlo simulations to show the usefulness of the findings. Finally, we apply the findings of the paper to some real data set.     

A Sarmanov family with beta and gamma marginal distributions: an application to the Bayes premium in a collective risk model

In this paper we firstly develop a Sarmanov–Lee bivariate family of distributions with the beta and gamma as marginal distributions. We obtain the linear correlation coefficient showing that, although it is not a strong family of correlation, it can be greater than the value of this coefficient in the Farlie–Gumbel–Morgenstern family. We also determine other measures for this family: the coefficient of median concordance and the relative entropy, which are analyzed by comparison with the case of independence. Secondly, we consider the problem of premium calculation in a Poisson–Lindley and exponential collective risk model, where the Sarmanov–Lee family is used as a structure function. We determine the collective and Bayes premiums whose values are analyzed when independence and dependence between the risk profiles are considered, obtaining that notable variations in premiums values are obtained even when low levels of correlation are considered.     

The inverse power Lindley distribution in the presence of left-censored data

In this study, classical and Bayesian inference methods are introduced to analyze lifetime data sets in the presence of left censoring considering two generalizations of the Lindley distribution: a first generalization proposed by Ghitany et al. [Power Lindley distribution and associated inference, Comput. Statist. Data Anal. 64 (2013), pp. 20–33], denoted as a power Lindley distribution and a second generalization proposed by Sharma et al. [The inverse Lindley distribution: A stress–strength reliability model with application to head and neck cancer data, J. Ind. Prod. Eng. 32 (2015), pp. 162–173], denoted as an inverse Lindley distribution. In our approach, we have used a distribution obtained from these two generalizations denoted as an inverse power Lindley distribution. A numerical illustration is presented considering a dataset of thyroglobulin levels present in a group of individuals with differentiated cancer of thyroid.     

Doubly robust empirical likelihood inference in covariate-missing data problems

Missing covariate data occurs often in regression analysis. We study methods for estimating the regression coefficients in an assumed conditional mean function when some covariates are completely observed but other covariates are missing for some subjects. We adopt the semiparametric perspective of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Amer Statist Assoc. 1994;89:846–866] on regression analyses with missing covariates, in which they pioneered the use of two working models, the working propensity score model and the working conditional score model. A recent approach to missing covariate data analysis is the empirical likelihood method of Qin et al. [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503], which effectively combines unbiased estimating equations. In this paper, we consider an alternative likelihood approach based on the full likelihood of the observed data. This full likelihood-based method enables us to generate estimators for the vector of the regression coefficients that are (a) asymptotically equivalent to those of Qin et al. [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503] when the working propensity score model is correctly specified, and (b) doubly robust, like the augmented inverse probability weighting (AIPW) estimators of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Am Statist Assoc. 1994;89:846–866]. Thus, the proposed full likelihood-based estimators improve on the efficiency of the AIPW estimators when the working propensity score model is correct but the working conditional score model is possibly incorrect, and also improve on the empirical likelihood estimators of Qin, Zhang and Leung [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503] when the reverse is true, that is, the working conditional score model is correct but the working propensity score model is possibly incorrect. In addition, we consider a regression method for estimation of the regression coefficients when the working conditional score model is correctly specified; the asymptotic variance of the resulting estimator is no greater than the semiparametric variance bound characterized by the theory of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Amer Statist Assoc. 1994;89:846–866]. Finally, we compare the finite-sample performance of various estimators in a simulation study.     

Bayes estimation of Gompertz distribution parameters and acceleration factor under partially accelerated life tests with type-I censoring

In this paper, the Bayesian approach is applied to the estimation problem in the case of step stress partially accelerated life tests with two stress levels and type-I censoring. Gompertz distribution is considered as a lifetime model. The posterior means and posterior variances are derived using the squared-error loss function. The Bayes estimates cannot be obtained in explicit forms. Approximate Bayes estimates are computed using the method of Lindley [D.V. Lindley, Approximate Bayesian methods, Trabajos Estadistica 31 (1980), pp. 223–237]. The advantage of this proposed method is shown. The approximate Bayes estimates obtained under the assumption of non-informative priors are compared with their maximum likelihood counterparts using Monte Carlo simulation.     

The smooth colonel and the reverend find common ground

A semiparametric regression estimator that exploits categorical (i.e., discrete-support) kernel functions is developed for a broad class of hierarchical models including the pooled regression estimator, the fixed-effects estimator familiar from panel data, and the varying coefficient estimator, among others. Separate shrinking is allowed for each coefficient. Regressors may be continuous or discrete. The estimator is motivated as an intuitive and appealing generalization of existing methods. It is then supported by demonstrating that it can be realized as a posterior mean in the Lindley and Smith (1972 Lindley, D. V., Smith, A. F. M. (1972). Bayes estimates for the linear model. Journal of the Royal Statistical Society 34:1–41. [Google Scholar]) framework. As a demonstration of the flexibility of the proposed approach, the model is extended to nonparametric hierarchical regression based on B-splines.     

Inference for a simultaneous linear partially explosive model with polynomial regression components of different degrees

This paper is concerned with the estimation of the coefficients of simultaneous partially explosive model with polynomial regression components of different degrees in its equations. Since the least squares method breaks down in this case, a three stage estimation procedure is suggested for obtaining CAN estimates of the coefficients.     

Empirical distribution function under heteroscedasticity

Neglecting heteroscedasticity of error terms may imply the wrong identification of a regression model (see appendix). Employment of (heteroscedasticity resistent) White's estimator of covariance matrix of estimates of regression coefficients may lead to the correct decision about the significance of individual explanatory variables under heteroscedasticity. However, White's estimator of covariance matrix was established for least squares (LS)-regression analysis (in the case when error terms are normally distributed, LS- and maximum likelihood (ML)-analysis coincide and hence then White's estimate of covariance matrix is available for ML-regression analysis, tool). To establish White's-type estimate for another estimator of regression coefficients requires Bahadur representation of the estimator in question, under heteroscedasticity of error terms. The derivation of Bahadur representation for other (robust) estimators requires some tools. As the key too proved to be a tight approximation of the empirical distribution function (d.f.) of residuals by the theoretical d.f. of the error terms of the regression model. We need the approximation to be uniform in the argument of d.f. as well as in regression coefficients. The present paper offers this approximation for the situation when the error terms are heteroscedastic.     

A new lifetime model with decreasing failure rate

In this paper, we introduce a new lifetime distribution by compounding exponential and Poisson–Lindley distributions, named the exponential Poisson–Lindley (EPL) distribution. A practical situation where the EPL distribution is most appropriate for modelling lifetime data than exponential–geometric, exponential–Poisson and exponential–logarithmic distributions is presented. We obtain the density and failure rate of the EPL distribution and properties such as mean lifetime, moments, order statistics and Rényi entropy. Furthermore, estimation by maximum likelihood and inference for large samples are discussed. The paper is motivated by two applications to real data sets and we hope that this model will be able to attract wider applicability in survival and reliability.     

Analysis of an experiment aimed at improving the reliability of transmission centre shafts

Smith (1991) presents a paper proposing the use of Weibull regression models to establish dependence of failure data (usually times) on covariates related to the design of the test specimens and test procedures. In his article Smith made the point that good experimental design was as important in reliability applications as elsewhere, and in view of the current interest in design inspired by Taguchi and others, we pay some attention in this article to that topic. A real case study from the Ford Motor Company is presented.Our main approach is to utilize suggestions in the literature for applying standard least squares techniques of experimental analysis even when there is likely to be nonnormal error, and censoring. This approach lacks theoretical justification, but its appeal is its simplicity and flexibility. For completeness we also include some analysis based on the proportional hazards model, and in an attempt to link back to Smith (1991), look at a Weibull regression model.     

Evaluating fit in functional data analysis using model embeddings

The author proposes a general method for evaluating the fit of a model for functional data. His approach consists of embedding the proposed model into a larger family of models, assuming the true process generating the data is within the larger family, and then computing a posterior distribution for the Kullback‐Leibler distance between the true and the proposed models. The technique is illustrated on biomechanical data reported by Ramsay, Flanagan & Wang (1995). It is developed in detail for hierarchical polynomial models such as those found in Lindley & Smith (1972), and is also generally applicable to longitudinal data analysis where polynomials are fit to many individuals.     

Approximate tolerance intervals for the discrete Poisson–Lindley distribution

The Poisson–Lindley distribution is a compound discrete distribution that can be used as an alternative to other discrete distributions, like the negative binomial. This paper develops approximate one-sided and equal-tailed two-sided tolerance intervals for the Poisson–Lindley distribution. Practical applications of the Poisson–Lindley distribution frequently involve large samples, thus we utilize large-sample Wald confidence intervals in the construction of our tolerance intervals. A coverage study is presented to demonstrate the efficacy of the proposed tolerance intervals. The tolerance intervals are also demonstrated using two real data sets. The R code developed for our discussion is briefly highlighted and included in the tolerance package.     

High-correlated residuals improved estimation in the high-dimensional SUR model

This article focuses on estimating regression coefficients in the high-dimensional seemingly unrelated regression model. When the number of equations exceeds that of the observations, both the maximum likelihood estimator and Zellner’s two-stage estimator do not exist. As an alternative, we propose a two-stage conditional expectation improved estimator. The new estimator is further improved by the high-correlated residuals, and the high correlation is determined by hypothesis testings. Simulations show that the new estimator outperforms the ordinary least-squares estimator in terms of mean square errors, especially when high-correlated residuals exist between the equations.     

Estimation of tobit-type models with individual specific effects

The aim of this paper is two-fold. First, we review recent estimators for censored regression and sample selection panel data models with unobservable individual specific effects, and show how the idea behind these estimators can be used to construct estimators for a variety of other Tobit-type models. The estimators presented in this paper are semiparametric, in the sense that they do not require the parametrization of the distribution of the unobservables. The second aim of the paper is to introduce a new class of estimators for the censored regression model. The advantage of the new estimators is that they can be applied under a stationarity assumption on the transitory error terms, which is weaker than the exchangeability assumption that is usually made in this literature. A similar generalization does not seem feasible for the estimators of the other models that are considered.     

Estimation of tobit-type models with individual specific effects

The aim of this paper is two-fold. First, we review recent estimators for censored regression and sample selection panel data models with unobservable individual specific effects, and show how the idea behind these estimators can be used to construct estimators for a variety of other Tobit-type models. The estimators presented in this paper are semiparametric, in the sense that they do not require the parametrization of the distribution of the unobservables. The second aim of the paper is to introduce a new class of estimators for the censored regression model. The advantage of the new estimators is that they can be applied under a stationarity assumption on the transitory error terms, which is weaker than the exchangeability assumption that is usually made in this literature. A similar generalization does not seem feasible for the estimators of the other models that are considered.     

Median regression for SUR models with the same explanatory variables in each equation

In this paper we introduce an interesting feature of the generalized least absolute deviations method for seemingly unrelated regression equations (SURE) models. Contrary to the collapse of generalized leasts-quares parameter estimations of SURE models to the ordinary least-squares estimations of the individual equations when the same regressors are common between all equations, the estimations of the proposed methodology are not identical to the least absolute deviations estimations of the individual equations. This is important since contrary to the least-squares methods, one can take advantage of efficiency gain due to cross-equation correlations even if the system includes the same regressors in each equation.     

Power-normal distribution

Recently, Gupta and Gupta [Analyzing skewed data by power-normal model, Test 17 (2008), pp. 197–210] proposed the power-normal distribution for which normal distribution is a special case. The power-normal distribution is a skewed distribution, whose support is the whole real line. Our main aim of this paper is to consider bivariate power-normal distribution, whose marginals are power-normal distributions. We obtain the proposed bivariate power-normal distribution from Clayton copula, and by making a suitable transformation in both the marginals. Lindley–Singpurwalla distribution also can be used to obtain the same distribution. Different properties of this new distribution have been investigated in detail. Two different estimators are proposed. One data analysis has been performed for illustrative purposes. Finally, we propose some generalizations to multivariate case also along the same line and discuss some of its properties.     

Exponentiated power Lindley–Poisson distribution: Properties and applications

A new four-parameter distribution called the exponentiated power Lindley–Poisson distribution which is an extension of the power Lindley and Lindley–Poisson distributions is introduced. Statistical properties of the distribution including the shapes of the density and hazard functions, moments, entropy measures, and distribution of order statistics are given. Maximum likelihood estimation technique is used to estimate the parameters. A simulation study is conducted to examine the bias, mean square error of the maximum likelihood estimators, and width of the confidence interval for each parameter. Finally, applications to real data sets are presented to illustrate the usefulness of the proposed distribution.