Análise bayesiana do modelo fatorial dinâmico para um vetor de séries temporais utilizando a distribuição t multivariada

The multivariate t models are symmetric and with heavier tail than the normal distribution, important feature in financial data. In this theses is presented the Bayesian estimation of a dynamic factor model, where the factors follow a multivariate autoregressive model, using multivariate t distribution. Since the multivariate t distribution is complex, it was represented in this work as a mix between a multivariate normal distribution and a square root of a chi-square distribution. This method allowed to define the posteriors. The inference on the parameters was made taking a sample of the posterior distribution, through the Gibbs Sampler. The convergence was verified through graphical analysis and the convergence tests Geweke (1992) and Raftery & Lewis (1992a). The method was applied in simulated data and in the indexes of the major stock exchanges in the world.

Propriedades e aspectos geométricos de estimadores tipo James-Stein e do estimador de Hartigan

The James-Stein estimator is a biased shrinkage estimator with uniformly smaller risk than the risk of the sample mean estimator for the mean of multivariate normal distribution, except in the one-dimensional or two-dimensional cases. In this work we have used more heuristic arguments and intensified the geometric treatment of the theory of James-Stein estimator. New type James-Stein shrinking estimators are proposed and the Mahalanobis metric used to address the James-Stein estimator. . To evaluate the performance of the estimator proposed, in relation to the sample mean estimator, we used the computer simulation by the Monte Carlo method by calculating the mean square error. The result indicates that the new estimator has better performance relative to the sample mean estimator.

Modelos duplo sigmoidais para o crescimento de frutos do cafeeiro

Coffee is one of the main products of Brazilian agriculture, the country is currently the largest producer and exporter. Knowing the growth pattern of a fruit can assist in the development of culture indicating for example, the times of increased fruit weight and its optimum harvest, essential to improve the management and quality of coffee. Some authors indicate that the growth curve of the coffee fruit has a double sigmoid shape. However, it consists of just a visual observation without exploring the use of regression models. The aims of this study were: i) determine if the growth pattern of the coffee fruit is really double sigmoidal; ii) to propose a new approach in weighted importance re-sampling to estimate the parameters of regression models and select the most suitable double sigmoidal model to describe the growth of coffee fruits; iii) to study the spatial distribution effect of the crop in the growth curve of coffee fruits. In the first article the aim was determine if the growth pattern of the coffee fruit is really double sigmoidal. The models double Gompertz and double Logistic showed significantly superior fit to models of simple sigmoid confirming that the standard of coffee fruits growth is really double sigmoidal. In the second article we propose to consider an approximation of the likelihood as the candidate distribution of the weighted importance resampling, aiming to facilitate the process of obtaining samples of marginal distributions of each parameter. This technique was effective since it provided parameters with practical interpretation and low computational effort, therefore, it can be used to estimate parameters of double sigmoidal growth curves. The nonlinear model double Logistic was the most appropriate to describe the growth curve of coffee fruits. In the third article aimed to verify the influence of different planting alignments and sun exposure faces in the fruits growth curve. A difference between the growth rates in the two stages of fruit development was identified, regardless the side. Although it has been proven differences in productivity and quality of coffee, there was no difference between the growth curves in the different planting alignments herein studied.

Análise de experimentos fatoriais de dois fatores com tratamentos adicionais

Experiments using factorial arrangement of treatments are rather common and useful in agricultural research. The main advantage is the possibility of testing many hypotheses, allowing broader conclusions over different factors, studied simultaneously. Factorial arrangements are used to compare levels of each factor (main effects), and also to verify if the differences among levels of a given factor are dependent on the levels of the other factors (interacions). In the analysis of data from factorial experiments, difficulty is increased when additional treatments are included. Inclusion of one or more additional treatments is a quite common practice, since such treatments are usually taken as reference or standard for evaluation and comparasion of the remaimng treatments, or aiming complementary information. This increase of difficulty is however low, compared to the advantages. As in literature there are few references about the statistical analysis of factorial experiments with additonal treatments, and given straightforward use in experimentation, the objective of this work was the presentation of approach for the use of factorial experiments with additonal treatments trough the analysis of some examples using the SAS® software, with the corresponding theoretical development, obtaining the system of normal equations, estimators of the parameters and variance of contrasts among two treatment means. It is suggested that additional treatments should be used with caution. The analysis of variance of such kind of experiment was presented using matrix notation.

Regressão Simplex aplicada a delineamentos de mistura e utilização do Algoritmo Boosting

In the composition of this work are present two parts. The first part contains the theory used. The second part contains the two articles. The first article examines two models of the class of generalized linear models for analyzing a mixture experiment, which studied the effect of different diets consist of fat, carbohydrate, and fiber on tumor expression in mammary glands of female rats, given by the ratio mice that had tumor expression in a particular diet. Mixture experiments are characterized by having the effect of collinearity and smaller sample size. In this sense, assuming normality for the answer to be maximized or minimized may be inadequate. Given this fact, the main characteristics of logistic regression and simplex models are addressed. The models were compared by the criteria of selection of models AIC, BIC and ICOMP, simulated envelope charts for residuals of adjusted models, odds ratios graphics and their respective confidence intervals for each mixture component. It was concluded that first article that the simplex regression model showed better quality of fit and narrowest confidence intervals for odds ratio. The second article presents the model Boosted Simplex Regression, the boosting version of the simplex regression model, as an alternative to increase the precision of confidence intervals for the odds ratio for each mixture component. For this, we used the Monte Carlo method for the construction of confidence intervals. Moreover, it is presented in an innovative way the envelope simulated chart for residuals of the adjusted model via boosting algorithm. It was concluded that the Boosted Simplex Regression model was adjusted successfully and confidence intervals for the odds ratio were accurate and lightly more precise than the its maximum likelihood version.