Methods of Estimation in Multiple Linear Regression: Application to Clinical Data

Nesse artigo, tem-se o interesse em avaliar diferentes estratégias de estimação de parâmetros para um modelo de regressão linear múltipla. Para a estimação dos parâmetros do modelo foram utilizados dados de um ensaio clínico em que o interesse foi verificar se o ensaio mecânico da propriedade de força máxima (EM-FM) está associada com a massa femoral, com o diâmetro femoral e com o grupo experimental de ratas ovariectomizadas da raça Rattus norvegicus albinus, variedade Wistar. Para a estimação dos parâmetros do modelo serão comparadas três metodologias: a metodologia clássica, baseada no método dos mínimos quadrados; a metodologia Bayesiana, baseada no teorema de Bayes; e o método Bootstrap, baseado em processos de reamostragem.

Using stochastic volatility models to analyse weekly ozone averages in Mexico City

In this paper we make use of some stochastic volatility models to analyse the behaviour of a weekly ozone average measurements series. The models considered here have been used previously in problems related to financial time series. Two models are considered and their parameters are estimated using a Bayesian approach based on Markov chain Monte Carlo (MCMC) methods. Both models are applied to the data provided by the monitoring network of the Metropolitan Area of Mexico City. The selection of the best model for that specific data set is performed using the Deviance Information Criterion and the Conditional Predictive Ordinate method.

Graphical models and point set matching

Point pattern matching in Euclidean Spaces is one of the fundamental problems in Pattern Recognition, having applications ranging from Computer Vision to Computational Chemistry. Whenever two complex patterns are encoded by two sets of points identifying their key features, their comparison can be seen as a point pattern matching problem. This work proposes a single approach to both exact and inexact point set matching in Euclidean Spaces of arbitrary dimension. In the case of exact matching, it is assured to find an optimal solution. For inexact matching (when noise is involved), experimental results confirm the validity of the approach. We start by regarding point pattern matching as a weighted graph matching problem. We then formulate the weighted graph matching problem as one of Bayesian inference in a probabilistic graphical model. By exploiting the existence of fundamental constraints in patterns embedded in Euclidean Spaces, we prove that for exact point set matching a simple graphical model is equivalent to the full model. It is possible to show that exact probabilistic inference in this simple model has polynomial time complexity with respect to the number of elements in the patterns to be matched. This gives rise to a technique that for exact matching provably finds a global optimum in polynomial time for any dimensionality of the underlying Euclidean Space. Computational experiments comparing this technique with well-known probabilistic relaxation labeling show significant performance improvement for inexact matching. The proposed approach is significantly more robust under augmentation of the sizes of the involved patterns. In the absence of noise, the results are always perfect.

The best are never normal : exploring the distribution of firm performance

Competitive Strategy literature predicts three different mechanisms of performance generation, thus distinguishing between firms that have competitive advantage, firms that have competitive disadvantage or firms that have neither. Nonetheless, previous works in the field have fitted a single normal distribution to model firm performance. Here, we develop a new approach that distinguishes among performance generating mechanisms and allows the identification of firms with competitive advantage or disadvantage. Theorizing on the positive feedback loops by which firms with competitive advantage have facilitated access to acquire new resources, we proposed a distribution we believe data on firm performance should follow. We illustrate our model by assessing its fit to data on firm performance, addressing its theoretical implications and comparing it to previous works.

Modelos dinâmicos e simulação estocástica

This paper presents new methodology for making Bayesian inference about dy~ o!s for exponential famiIy observations. The approach is simulation-based _~t> use of ~vlarkov chain Monte Carlo techniques. A yletropolis-Hastings i:U~UnLlllll 1::; combined with the Gibbs sampler in repeated use of an adjusted version of normal dynamic linear models. Different alternative schemes are derived and compared. The approach is fully Bayesian in obtaining posterior samples for state parameters and unknown hyperparameters. Illustrations to real data sets with sparse counts and missing values are presented. Extensions to accommodate for general distributions for observations and disturbances. intervention. non-linear models and rnultivariate time series are outlined.

Simulation-based smoothing and filtering in factor stochastic volatility models : two econometric applications

In this article we use factor models to describe a certain class of covariance structure for financiaI time series models. More specifical1y, we concentrate on situations where the factor variances are modeled by a multivariate stochastic volatility structure. We build on previous work by allowing the factor loadings, in the factor mo deI structure, to have a time-varying structure and to capture changes in asset weights over time motivated by applications with multi pIe time series of daily exchange rates. We explore and discuss potential extensions to the models exposed here in the prediction area. This discussion leads to open issues on real time implementation and natural model comparisons.

Simulation -based smoothing and filtering in factor stochastic volatility models : two econometric applications (July-2001)

The past decade has wítenessed a series of (well accepted and defined) financial crises periods in the world economy. Most of these events aI,"e country specific and eventually spreaded out across neighbor countries, with the concept of vicinity extrapolating the geographic maps and entering the contagion maps. Unfortunately, what contagion represents and how to measure it are still unanswered questions. In this article we measure the transmission of shocks by cross-market correlation\ coefficients following Forbes and Rigobon's (2000) notion of shift-contagion,. Our main contribution relies upon the use of traditional factor model techniques combined with stochastic volatility mo deIs to study the dependence among Latin American stock price indexes and the North American indexo More specifically, we concentrate on situations where the factor variances are modeled by a multivariate stochastic volatility structure. From a theoretical perspective, we improve currently available methodology by allowing the factor loadings, in the factor model structure, to have a time-varying structure and to capture changes in the series' weights over time. By doing this, we believe that changes and interventions experienced by those five countries are well accommodated by our models which learns and adapts reasonably fast to those economic and idiosyncratic shocks. We empirically show that the time varying covariance structure can be modeled by one or two common factors and that some sort of contagion is present in most of the series' covariances during periods of economical instability, or crisis. Open issues on real time implementation and natural model comparisons are thoroughly discussed.

Efeitos genéticos e ambientais sobre o perímetro escrotal de animais da raça Caracu

Neste trabalho foram avaliados os efeitos de idade (IDS) e peso aos 15 meses (P15) sobre medidas de perímetro escrotal ao sobreano de 1.892 machos da raça Caracu e estimada a herdabilidade desta característica. Utilizaram-se dois modelos de análise: em um dos modelos, foram incluídos o efeito de grupo de contemporâneos (GC) e, como covariável, a idade ao sobreano (efeitos linear e quadrático); e, no outro, o efeito de GC e, como covariável, o peso ao sobreano (efeitos linear e quadrático). A idade não teve efeito significativo no perímetro escrotal, o que está relacionado à homogeneidade dos animais nos grupos de contemporâneos, enquanto os efeitos linear e quadrático de P15 foram significativos sobre o perímetro escrotal, indicando grande influência do peso sobre a variação desta característica. A herdabilidade do perímetro escrotal foi estimada pelo método Bayesiano utilizando-se um modelo animal. O modelo incluiu os efeitos de GC e P15 (linear e quadrático) e os efeitos genéticos aditivo direto e residual. O valor médio estimado de herdabilidade do PE ao sobreano foi 0,38 e comprova que esta característica pode ser utilizada como critério de seleção para precocidade sexual em programas de melhoramento genético de animais da raça Caracu.

Genotype by environment interaction in Nelore cattle from five Brazilian states

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Testicular traits as selection criteria for young Nellore bulls

The objective of this study was to evaluate the possible use of biometric testicular traits as selection criteria for young Nellore bulls using Bayesian inference to estimate heritability coefficients and genetic correlations. Multitrait analysis was performed including 17,211 records of scrotal circumference obtained during andrological assessment (SCAND) and 15,313 records of testicular volume and shape. In addition, 50,809 records of scrotal circumference at 18 mo (SC18), used as an anchor trait, were analyzed. The (co) variance components and breeding values were estimated by Gibbs sampling using the Gibbs2F90 program under an animal model that included contemporary groups as fixed effects, age of the animal as a linear covariate, and direct additive genetic effects as random effects. Heritabilities of 0.42, 0.43, 0.31, 0.20, 0.04, 0.16, 0.15, and 0.10 were obtained for SC18, SCAND, testicular volume, testicular shape, minor defects, major defects, total defects, and satisfactory andrological evaluation, respectively. The genetic correlations between SC18 and the other traits were 0.84 (SCAND), 0.75 (testicular shape), 0.44 (testicular volume), -0.23 (minor defects), -0.16 (major defects), -0.24 (total defects), and 0.56 (satisfactory andrological evaluation). Genetic correlations of 0.94 and 0.52 were obtained between SCAND and testicular volume and shape, respectively, and of 0.52 between testicular volume and testicular shape. In addition to favorable genetic parameter estimates, SC18 was found to be the most advantageous testicular trait due to its easy measurement before andrological assessment of the animals, even though the utilization of biometric testicular traits as selection criteria was also found to be possible. In conclusion, SC18 and biometric testicular traits can be adopted as a selection criterion to improve the fertility of young Nellore bulls.

Genetic associations of visual scores with subsequent rebreeding and days to first calving in Nellore cattle

Records of Nellore animals born from 1990 to 2006 were used to estimate genetic correlations of visual scores at yearling (conformation, C; finishing precocity, P; and muscling, M) with primiparous subsequent rebreeding (SR) and days to first calving (DC), because the magnitude of these associations is still unknown. Genetic parameters were estimated by multiple-traits Bayesian analysis, using a nonlinear (threshold) animal models for visual scores and SR and a linear animal models for weaning weight (WW) and DC. WW was included in the analysis to account for the effects of sequential selection. The posterior means of heritabilities estimated for C, P, M, SR and DC were 0.24 +/- 0.01, 0.31 +/- 0.01, 0.30 +/- 0.01, 0.18 +/- 0.02 and 0.06 +/- 0.02, respectively. The posterior means of genetic correlations estimated between SR and visual scores were low and positive, with values of 0.09 +/- 0.02 (C), 0.19 +/- 0.03 (P) and 0.18 +/- 0.05 (M). on the other hand, negative genetic correlations were found between DC and C (-0.11 +/- 0.09), P (-0.19 +/- 0.09) and M (-0.16 +/- 0.09). The primiparous rebreeding trait has genetic variability in Nellore cattle. The genetic correlations between visual scores, and SR and DC were low and favourable. The genetic changes in C, P and M were 0.02, 0.03 and 0.03/year, respectively. For SR and DC, genetic trends were 0.01/year and -0.01 days/year, respectively, indicating that the increase in genetic merit for reproductive traits was small over time. Direct selection for visual scores together with female reproductive traits is recommended to increase the fertility of beef cows.

Genetic associations of sexual precocity with growth traits and visual scores of conformation, finishing, and muscling in Nelore cattle

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Genetic parameters and relationships between heifers rebreeding and hip height in Nellore cattle

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Associação genética da prenhez aos 16 meses com o peso à desmama e o ganho de peso em animais da raça Nelore

Objetivou-se verificar a possibilidade de utilização da prenhez de novilhas aos 16 meses (Pr16) como critério de seleção e as possíveis associações genéticas entre prenhez em novilhas aos 16 meses e o peso à desmama (PD) e o ganho de peso médio da desmama ao sobreano (GP). Foram realizadas análises uni e bicaracterísticas para estimação dos componentes de co-variância, empregando-se um modelo animal linear para peso à desmama e ganho de peso da desmama ao sobreano e não-linear para Pr16. A estimação dos componentes de variância e da predição dos valores genéticos dos animais foi realizada por Inferência Bayesiana. Distribuições flat foram utilizadas para todos os componentes de co-variância. As estimativas de herdabilidade direta para Pr16, PD e GP foram 0,50; 0,24 e 0,15, respectivamente, e a estimativa de herdabilidade materna para o PD, de 0,07. As correlações genéticas foram -0,25 e 0,09 entre Pr16, PD e GP, respectivamente, e a correlação genética entre Pr16 e o efeito genético materno do PD, de 0,29. A herdabilidade da prenhez aos 16 meses indica que essa característica pode ser utilizada como critério de seleção. As correlações genéticas estimadas indicam que a seleção por animais mais pesados à desmama, a longo prazo, pode diminuir a ocorrência de prenhez aos 16 meses de idade. Além disso, a seleção para maior habilidade materna favorece a seleção de animais mais precoces. No entanto, a seleção para ganho de peso da desmama ao sobreano não leva a mudanças genéticas na precocidade sexual em fêmeas.

Características reprodutivas e suas associações com outras características de importância econômica na raça Nelore

Objetivou-se com este trabalho estimar as herdabilidades (h²) e as correlações genéticas (r g) entre idade ao primeiro parto (IPP) e primeiro intervalo de partos (PIEP) e outras características como peso (PS) ao ano (A) e ao sobreano (S), altura do posterior (ALT) e perímetro escrotal (PE450) em animais da raça Nelore. Os parâmetros genéticos foram estimados em uma análise multicaracterística por modelo animal, utilizando-se a inferência bayesiana via algoritmo de Gibbs Sampling. Os parâmetros genéticos estimados sugerem a existência de variabilidade genética para IPP (h² = 0,26), sendo que a seleção para a diminuição da IPP de fêmeas Nelore deve responder à seleção individual, sem causar antagonismo do valor genético dos animais para PS (r g = -0,22 (A) e -0,44 (S)) e PE450 (r g = 0,02). A seleção para a diminuição da IPP, no longo prazo, pode levar a um aumento da ALT dos animais, embora essa associação seja relativamente baixa (-0,35). A estimativa de herdabilidade a posteriori para a característica PIEP foi baixa, 0,11±0,03. As r g entre PIEP e as demais características estudadas indicam que a seleção para essas características de crescimento não afetará o PIEP.