Likelihood ratio tests for variance components in linear mixed models

Although the asymptotic distributions of the likelihood ratio for testing hypotheses of null variance components in linear mixed models derived by Stram and Lee [1994. Variance components testing in longitudinal mixed effects model. Biometrics 50, 1171-1177] are valid, their proof is based on the work of Self and Liang [1987. Asymptotic properties of maximum likelihood estimators and likelihood tests under nonstandard conditions. J. Amer. Statist. Assoc. 82, 605-610] which requires identically distributed random variables, an assumption not always valid in longitudinal data problems. We use the less restrictive results of Vu and Zhou [1997. Generalization of likelihood ratio tests under nonstandard conditions. Ann. Statist. 25, 897-916] to prove that the proposed mixture of chi-squared distributions is the actual asymptotic distribution of such likelihood ratios used as test statistics for null variance components in models with one or two random effects. We also consider a limited simulation study to evaluate the appropriateness of the asymptotic distribution of such likelihood ratios in moderately sized samples. (C) 2008 Elsevier B.V. All rights reserved.

Hypothesis testing in an errors-in-variables model with heteroscedastic measurement errors

In many epidemiological studies it is common to resort to regression models relating incidence of a disease and its risk factors. The main goal of this paper is to consider inference on such models with error-prone observations and variances of the measurement errors changing across observations. We suppose that the observations follow a bivariate normal distribution and the measurement errors are normally distributed. Aggregate data allow the estimation of the error variances. Maximum likelihood estimates are computed numerically via the EM algorithm. Consistent estimation of the asymptotic variance of the maximum likelihood estimators is also discussed. Test statistics are proposed for testing hypotheses of interest. Further, we implement a simple graphical device that enables an assessment of the model`s goodness of fit. Results of simulations concerning the properties of the test statistics are reported. The approach is illustrated with data from the WHO MONICA Project on cardiovascular disease. Copyright (C) 2008 John Wiley & Sons, Ltd.

Generalized Birnbaum-Saunders distributions applied to air pollutant concentration

The generalized Birnbaum-Saunders (GBS) distribution is a new class of positively skewed models with lighter and heavier tails than the traditional Birnbaum-Saunders (BS) distribution, which is largely applied to study lifetimes. However, the theoretical argument and the interesting properties of the GBS model have made its application possible beyond the lifetime analysis. The aim of this paper is to present the GBS distribution as a useful model for describing pollution data and deriving its positive and negative moments. Based on these moments, we develop estimation and goodness-of-fit methods. Also, some properties of the proposed estimators useful for developing asymptotic inference are presented. Finally, an application with real data from Environmental Sciences is given to illustrate the methodology developed. This example shows that the empirical fit of the GBS distribution to the data is very good. Thus, the GBS model is appropriate for describing air pollutant concentration data, which produces better results than the lognormal model when the administrative target is determined for abating air pollution. Copyright (c) 2007 John Wiley & Sons, Ltd.

DIAGNOSTIC TECHNIQUES OF LOCAL INFLUENCE IN SPATIAL ANALYSIS OF SOYBEAN YIELD

Modeling of spatial dependence structure, concerning geoestatistics approach, is an indispensable tool for fixing parameters that define this structure, applied on interpolation of values in places that are not sampled, by kriging techniques. However, the estimation of parameters can be greatly affected by the presence of atypical observations on sampled data. Thus, this trial aimed at using diagnostics techniques of local influence in spatial linear Gaussians models, applied at geoestatistics in order to evaluate sensitivity of maximum likelihood estimators and restrict maximum likelihood to small perturbations in these data. So, studies with simulated and experimental data were performed. Those results, obtained from the study of real data, allowed us to conclude that the presence of atypical values among the sampled data can have a strong influence on thematic maps, changing, therefore, the spatial dependence. The application of diagnostics techniques of local influence should be part of any geoestatistic analysis, ensuring that the information from thematic maps has better quality and can be used with greater security by farmers.

Improved maximum-likelihood estimation in a regression model with general parametrization

We analyse the finite-sample behaviour of two second-order bias-corrected alternatives to the maximum-likelihood estimator of the parameters in a multivariate normal regression model with general parametrization proposed by Patriota and Lemonte [A. G. Patriota and A. J. Lemonte, Bias correction in a multivariate regression model with genereal parameterization, Stat. Prob. Lett. 79 (2009), pp. 1655-1662]. The two finite-sample corrections we consider are the conventional second-order bias-corrected estimator and the bootstrap bias correction. We present the numerical results comparing the performance of these estimators. Our results reveal that analytical bias correction outperforms numerical bias corrections obtained from bootstrapping schemes.

Covariance matrix formula for Birnbaum-Saunders regression models

The Birnbaum-Saunders regression model is commonly used in reliability studies. We derive a simple matrix formula for second-order covariances of maximum-likelihood estimators in this class of models. The formula is quite suitable for computer implementation, since it involves only simple operations on matrices and vectors. Some simulation results show that the second-order covariances can be quite pronounced in small to moderate sample sizes. We also present empirical applications.

Asymptotic skewness in Birnbaum-Saunders nonlinear regression models

The family of distributions proposed by Birnbaum and Saunders (1969) can be used to model lifetime data and it is widely applicable to model failure times of fatiguing materials. We give a simple matrix formula of order n(-1/2), where n is the sample size, for the skewness of the distributions of the maximum likelihood estimates of the parameters in Birnbaum-Saunders nonlinear regression models, recently introduced by Lemonte and Cordeiro (2009). The formula is quite suitable for computer implementation, since it involves only simple operations on matrices and vectors, in order to obtain closed-form skewness in a wide range of nonlinear regression models. Empirical and real applications are analyzed and discussed. (C) 2010 Elsevier B.V. All rights reserved.

Birnbaum-Saunders nonlinear regression models

We introduce, for the first time, a new class of Birnbaum-Saunders nonlinear regression models potentially useful in lifetime data analysis. The class generalizes the regression model described by Rieck and Nedelman [Rieck, J.R., Nedelman, J.R., 1991. A log-linear model for the Birnbaum-Saunders distribution. Technometrics 33, 51-60]. We discuss maximum-likelihood estimation for the parameters of the model, and derive closed-form expressions for the second-order biases of these estimates. Our formulae are easily computed as ordinary linear regressions and are then used to define bias corrected maximum-likelihood estimates. Some simulation results show that the bias correction scheme yields nearly unbiased estimates without increasing the mean squared errors. Two empirical applications are analysed and discussed. Crown Copyright (C) 2009 Published by Elsevier B.V. All rights reserved.

Bias correction in a multivariate normal regression model with general parameterization

This paper derives the second-order biases Of maximum likelihood estimates from a multivariate normal model where the mean vector and the covariance matrix have parameters in common. We show that the second order bias can always be obtained by means of ordinary weighted least-squares regressions. We conduct simulation studies which indicate that the bias correction scheme yields nearly unbiased estimators. (C) 2009 Elsevier B.V. All rights reserved.

Design and assessment of a large commercial Photovoltaic System in Barbados

One of the main aims of this thesis is to design an optimized commercial Photovoltaic (PV) system in Barbados from several variables such as racking type, module type and inverter type based on practicality, technical performance as well as financial returns to the client. Detailed simulations are done in PVSYST and financial models are used to compare different systems and their viability. Once the preeminent system is determined from a financial and performance perspective a detailed design is done using PVSYST and AutoCAD to design the most optimal PV system for the customer. In doing so, suitable engineering drawings are generated which are detailed enough for construction of the system. Detailed cost with quotes from relevant manufacturers, suppliers and estimators become instrumental in determining Balance of System Costs in addition to total project cost. The final simulated system is suggested with a PV capacity of 425kW and an inverter output of 300kW resulting in an array oversizing of 1.42. The PV system has a weighted Performance Ratio of 77 %, a specific yield of 1467 kWh/kWp and a projected annual production of 624 MWh/yr. This system is estimated to offset approximately 28 % of Carlton’s electrical load annually. Over the course of 20 years the PV system is projected to produce electricity at a cost of $0.201USD/kWh which is significantly lower than the $0.35 USD/kWh paid to the utility at the time of writing this thesis. Due to the high cost of electricity on the island, an attractive Feed-In-Tariff is not necessary to warrant the installation of a commercial System which over a lifetime which produces electricity at less than 60% of the cost to the user purchasing electricity from the utility. A simple payback period of 5.4 years, a return on investment of 17 % without incentives, in addition to an estimated diversion of 6840 barrels of oil or 2168 tonnes of CO2 further provides compelling justification for the installation of a commercial Photovoltaic System not only on Carlton A-1 Supermarket, but also island wide as well as regionally where most electricity supplies are from imported fossil fuels.

Statistical bound of genetic solutions to quadratic assignment problems

Quadratic assignment problems (QAPs) are commonly solved by heuristic methods, where the optimum is sought iteratively. Heuristics are known to provide good solutions but the quality of the solutions, i.e., the confidence interval of the solution is unknown. This paper uses statistical optimum estimation techniques (SOETs) to assess the quality of Genetic algorithm solutions for QAPs. We examine the functioning of different SOETs regarding biasness, coverage rate and length of interval, and then we compare the SOET lower bound with deterministic ones. The commonly used deterministic bounds are confined to only a few algorithms. We show that, the Jackknife estimators have better performance than Weibull estimators, and when the number of heuristic solutions is as large as 100, higher order JK-estimators perform better than lower order ones. Compared with the deterministic bounds, the SOET lower bound performs significantly better than most deterministic lower bounds and is comparable with the best deterministic ones. 

On statistical bounds of heuristic solutions to location problems

Solutions to combinatorial optimization problems, such as problems of locating facilities, frequently rely on heuristics to minimize the objective function. The optimum is sought iteratively and a criterion is needed to decide when the procedure (almost) attains it. Pre-setting the number of iterations dominates in OR applications, which implies that the quality of the solution cannot be ascertained. A small, almost dormant, branch of the literature suggests using statistical principles to estimate the minimum and its bounds as a tool to decide upon stopping and evaluating the quality of the solution. In this paper we examine the functioning of statistical bounds obtained from four different estimators by using simulated annealing on p-median test problems taken from Beasley’s OR-library. We find the Weibull estimator and the 2nd order Jackknife estimator preferable and the requirement of sample size to be about 10 being much less than the current recommendation. However, reliable statistical bounds are found to depend critically on a sample of heuristic solutions of high quality and we give a simple statistic useful for checking the quality. We end the paper with an illustration on using statistical bounds in a problem of locating some 70 distribution centers of the Swedish Post in one Swedish region. 

A finite sample improvement of the fixed effects estimator : applied to technical infficiency

The FE ('fixed effects') estimator of technical inefficiency performs poorly when N ('number of firms') is large and T ('number of time observations') is small. We propose estimators of both the firm effects and the inefficiencies, which have small sample gains compared to the traditional FE estimator. The estimators are based on nonparametric kernel regression of unordered variables, which includes the FE estimator as a special case. In terms of global conditional MSE ('mean square error') criterions, it is proved that there are kernel estimators which are efficient to the FE estimators of firm effects and inefficiencies, in finite samples. Monte Carlo simulations supports our theoretical findings and in an empirical example it is shown how the traditional FE estimator and the proposed kernel FE estimator lead to very different conclusions about inefficiency of Indonesian rice farmers.

Spatial Inequality in Chile

Despite success in reducing poverty over the last twenty years, inequality in Chile has remained virtually unchanged, making Chile one of the least equal countries in the world. High levels of inequality have been shown to hamper further reductions in poverty as well as economic growth and local inequality has been shown to affect such outcomes as violence and health. The study of inequality at the local level is thus crucial for understanding the economic well-being of a country. Local measures of inequality have been difficult to obtain, but recent theoretical advances have enabled the combination of survey and census data to obtain estimators of inequality that are robust at disaggregated geographic levels. In this paper, we employ this methodology to produce consistent estimators of inequality for every county in Chile. We find a great deal of variation in inequality, with county-level Gini coefficients ranging from 0.41 to 0.63.

Estimação de regressões aditivas via backfitting e integração marginal: performance em pequenas amostras

Nesta dissertação realizou-se um experimento de Monte Carlo para re- velar algumas características das distribuições em amostras finitas dos estimadores Backfitting (B) e de Integração Marginal(MI) para uma regressão aditiva bivariada. Está-se particularmente interessado em fornecer alguma evidência de como os diferentes métodos de seleção da janela hn, tais co- mo os métodos plug-in, impactam as propriedades em pequenas amostras dos estimadores. Está-se interessado, também, em fornecer evidência do comportamento de diferentes estimadores de hn relativamente a seqüência ótima de hn que minimiza uma função perda escolhida. O impacto de ignorar a dependência entre os regressores na estimação da janela é tam- bém investigado. Esta é uma prática comum e deve ter impacto sobre o desempenho dos estimadores. Além disso, não há nenhuma rotina atual- mente disponível nos pacotes estatísticos/econométricos para a estimação de regressões aditivas via os métodos de Backfitting e Integração Marginal. É um dos objetivos a criação de rotinas em Gauss para a implementação prática destes estimadores. Por fim, diferentemente do que ocorre atual- mente, quando a utilização dos estimadores-B e MI é feita de maneira completamente ad-hoc, há o objetivo de fornecer a usuários informação que permita uma escolha mais objetiva de qual estimador usar quando se está trabalhando com uma amostra finita.