Acessos desde 01/11/2007:
Trabalhos Aprovados
Atenção: As comunicações foram provisoriamente aceitas para apresentação no 18o SINAPE. Apenas os trabalhos em que pelo menos um dos autores tiver, até 16/06, pago o boleto de inscrição do congresso terão a aceitação definitiva, os demais não poderão ser apresentados no evento. Verifique se o seu trabalho satisfaz esse requisito.
Filtrar por categoria/formato/palavra:
| Formato | Título | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Oral | Bayesian Analysis from Saturated Designs Métodos Bayesianos
Saturated designs are very useful for screening factors and in experiments where the observations are very difficult or expensive to obtain. Due to lack of degrees of freedom available to make inferences about the parameters in the model, the frequentist approach is extremely limited, so the Bayesian approach is a good way to analyse data from saturated designs. However, because of the reliance on prior information, Bayesian methods must be used carefully. Conjugate priors are shown to be somewhat inflexible, whereas priors using finite mixtures of densities yield more natural posterior densities. Also shown is that non-conjugate priors, with independence between the effect parameters and the variance, can be useful. | ||||||||
| Pôster | Bayesian Analysis to Correct False-negative Errors in Capture-recapture data Métodos Bayesianos
Population size estimation using capture-recapture models for photo-identification data may present some problems related to the identifiability of the individuals. As photo-id data used for animal recognition is based on natural markings, such as scars, colour pattern, or any trait related to species characteristics that can be used to distinguish the individuals, it could be the case that identification of a poorly marked individual may be overlooked. Consider, for example, a particular poorly marked individual that has already been captured in a previous sampling occasion. If that animal is photographed on a future occasion, but the analyst fails to recognize the correct match, two possibilities of mismatch may occur: (i) the photo may be associated to another individual already included in the data set, leading to a "false positive error", or (ii) the analyst can consider the photo as being from a new individual, resulting in a "false negative error", and, in this case, a new line in the data set will be created. Abundance estimates based on data which incorporate false-negative errors tend to be positively biased. | ||||||||
| Pôster | Bayesian and Maximum Likelihood Estimation Methods for State Space Models with Markovian Regime Switching Using Wavelets Séries Temporais
We propose a state space model with regime switching, where each regime is associated with different values for the parameters, regime transitions define a Markovian process, and transition probabilities are modelled using the logistic link function and wavelets. To evaluate the state variables and regime probabilities, the Kalman filter and a probability filter procedure conditional on each possible regime at each present and previous instants are used (Kim and Nelson, 1999). First, the maximum likelihood (ML) estimation is implemented using the EM-algorithm due to the presence of non-observable variables. In order to obtain standard errors of ML estimators, bootstrap for state space models (Stoffer and Wall, 1991) was modified including the simulation of non-observed regimes using smoothed probabilities. In the Bayesian approach, the estimation is feasible using the Gibbs sampling method, which provides the posterior distribution and, consequently, estimates and standard deviations. The main goal is to compare the estimatives and their standard errors using both estimation methods. The estimation method is illustrated with simulated data and with the United States monthly seasonally adjusted industrial production index from January 1960 to January 1995. | ||||||||
| Pôster | Bayesian Density Estimation Using Skew Student-t-Normal Mixtures Estatística em Ciências Sociais Aplicadas (Administração, Economia, Sociologia, Psicologia, etc.)
We present a Bayesian approach for modeling heterogeneous data and estimate multimodal densities using mixtures of Skew Student-t-Normal distributions (Gómez et al.; Environmetrics 2007; 18:395–407). A stochastic representation that is useful for implementing a MCMC Gibbs-type algorithm and results about existence of posterior moments are obtained. Data sets concerning the Gross Domestic Product per capita (Human Development Report) and body mass index (National Health and Nutrition Examination Survey), previously studied in the related literature, are analyzed. | ||||||||
| Oral | Bayesian Inference for the Skew Student-t-Normal Model Estatística em Meio Ambiente
We present a Bayesian approach for inference using the skew Student-t-normal model (G´omez et al., 2007), presenting a stochastic representation for it and proving that, even when the prior for the location parameter is not proper, the posterior distribution is proper and some posterior moments are finite. A MCMC sampler based on the representation is obtained, and a comparison involving different prior specifications is made. A real data set related to nickel concentration is analyzed to illustrate the methodology developed. Results seem to indicate Bayesian improvement over maximum likelihood large sample estimation. | ||||||||
| Pôster | Bayesian Joint Modelling of the Mean and Covariance Structures for Normal Longitudinal Data Modelos de Regressão
Researchers in many fields have been interested in the different modelling approaches for the mean and covariance structures in the context of longitudinal data. Most of these parametric approaches have concentrated on normal linear models and there have been only a few proposals within the Bayesian framework. In this paper, we consider the general antedependence model, and jointly model the mean and covariance structures, estimating the mean and autoregressive parameters, and the innovation variances in the longitudinal data context. We initially propose a new and computationally efficient classic estimation algorithm based on the Fisher scoring algorithm to obtain the estimators of the parameters. In addition, we also propose a Bayesian methodology based on the Gibbs Sampling, properly adapted for longitudinal data analysis, and that considers linear mean structures and unrestricted covariance structures for normal longitudinal data. Finally, we illustrate the proposed methodology and study its relative strengths and weaknesses when compared to other proposed methods by analyzing two examples, the race and the cattle data sets. Key Words: Bayes estimation, Fisher scoring, antedependence models, Gibbs sampling. | ||||||||
| Oral | Bayesian modelling of fnancial returns: a relationship between volatility and trading volume Econometria, Atuária e Finanças
The modified mixture model with Markov Switching volatility specifcation is introduced to analyze the relationship between stock return volatility and trading volume. We propose to construct an algorithm based on Markov Chain Monte Carlo (MCMC) simulation methods to estimate all the parameters in the model using the Bayesian approach. The series of returns and trading volume of the IBM stock will be analyzed. keywords: Stochastic volatility, Non linear and non Gaussian State Space Models, Markov process of rst order, Markov Chain Monte Carlo | ||||||||
| Pôster | BIOENGENHARIA DE SOLOS: ANÁLISE DE EROSÃO DO RIO PARAMOPAMA Estatística em Agronomia e Biologia
GROUND BIOENGINEERING: EROSION ANALYSIS IN PARAMOPAMA RIVER One of the serious problems from degradation environmental, aggravated or not by action human, is the erosion on the margins of rivers. Objectify regenerate degraded regions was realized a study of erosion contention margins of the Paramopama river, one of the affluent of the river Vaza-Barris located into the town of São Cristóvão, on coastal regional of the south of the state of Sergipe. By using techniques of Bioengenharia of Soils associate in experiment of subdivide parcels about to verification of the development from the species on the margins of the Paramopama river. Trying defining what species best adapted in affiliation to each treatment, determining the type of treatment is the best indicative about to the screening from erosion on the margins of rivers with like characteristics of the Paramopama. Complete what the treatment with the Syntemax 400BF with declivities is the of major cover vegetable average in relation to order of productivity mean from the species was such established: Calopogonium muconoides, Cajanus cajan, Crotalaria spectabilis, Vetiveria zizanoides. The best treatment specie was observer on affiliation among the treatment with Syntemax 400BF with declividade with the species Calopogonium muconoides. Keys words: Bioengenharia , Erosion , Subdivide parcels, Tukey’s test. | ||||||||
| Pôster | Bootstrap Confidence Intervals for Recurrent Event Data: A Simulation Study Estatística em Engenharia e Ciências Exatas
Experiments related to recurrent events provide information about the number of events, time to their ocurrence and their costs. Nelson (1995) presents a methodology to obtain confidence intervals for the cost and the number of cumulative events. Apart from this, it is possible to construct confidence intervals via computer-intensive methods, where the bootstrap is a particular case. In this work we present these two procedures. One of the advantages of the methodology presented in this work is the possibility for its application in several areas and its easy computational implementation. An example from engineering illustrates the methodology. Keywords: recurrent events, bootstrap, asymptotic theory, coverage probability. | ||||||||
| Pôster | BOOTSTRAPPING FOR ESTIMATION OF BIASES AND VARIANCES AT THE RESOURCE SELECTION PROBABILITY FUNCTION (RSPF) Estatística em Meio Ambiente
Resource selection functions are used for quantify how animals are selective in the use of the habitat period or food, Strickland and McDonald (2006). A resource selection probability function (RSPF) can be estimated if N, the total number of units in the population, and n1 the total number of used units in the study period are both known and small. An approximation the RSPF was then be estimated using any standard program for logistic regression standard but the variances of the estimates of the parameters are too small.\r\n\r\nThree methods of bootstrap sampling, parametric, non-parametric and a modified parametric method are proposed for the estimation of variances, with a discussion about the limitations of logistic regression for estimating RSPF. This method for estimating the RSPF described here has potential applications in medicine, ecology and other areas.\r\n\r\n\r\nKeywords: RSPFs, RSPF, Bootstrap.\r\n | ||||||||
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 | |||||||||