A novel MAP-MRF approach for multispectral image contextual classification using combination of suboptimal iterative algorithms

In this paper we present a novel approach for multispectral image contextual classification by combining iterative combinatorial optimization algorithms. The pixel-wise decision rule is defined using a Bayesian approach to combine two MRF models: a Gaussian Markov Random Field (GMRF) for the observations (likelihood) and a Potts model for the a priori knowledge, to regularize the solution in the presence of noisy data. Hence, the classification problem is stated according to a Maximum a Posteriori (MAP) framework. In order to approximate the MAP solution we apply several combinatorial optimization methods using multiple simultaneous initializations, making the solution less sensitive to the initial conditions and reducing both computational cost and time in comparison to Simulated Annealing, often unfeasible in many real image processing applications. Markov Random Field model parameters are estimated by Maximum Pseudo-Likelihood (MPL) approach, avoiding manual adjustments in the choice of the regularization parameters. Asymptotic evaluations assess the accuracy of the proposed parameter estimation procedure. To test and evaluate the proposed classification method, we adopt metrics for quantitative performance assessment (Cohen`s Kappa coefficient), allowing a robust and accurate statistical analysis. The obtained results clearly show that combining sub-optimal contextual algorithms significantly improves the classification performance, indicating the effectiveness of the proposed methodology. (C) 2010 Elsevier B.V. All rights reserved.

Indirect likelihood inference

Given a sample from a fully specified parametric model, let Zn be a given finite-dimensional statistic - for example, an initial estimator or a set of sample moments. We propose to (re-)estimate the parameters of the model by maximizing the likelihood of Zn. We call this the maximum indirect likelihood (MIL) estimator. We also propose a computationally tractable Bayesian version of the estimator which we refer to as a Bayesian Indirect Likelihood (BIL) estimator. In most cases, the density of the statistic will be of unknown form, and we develop simulated versions of the MIL and BIL estimators. We show that the indirect likelihood estimators are consistent and asymptotically normally distributed, with the same asymptotic variance as that of the corresponding efficient two-step GMM estimator based on the same statistic. However, our likelihood-based estimators, by taking into account the full finite-sample distribution of the statistic, are higher order efficient relative to GMM-type estimators. Furthermore, in many cases they enjoy a bias reduction property similar to that of the indirect inference estimator. Monte Carlo results for a number of applications including dynamic and nonlinear panel data models, a structural auction model and two DSGE models show that the proposed estimators indeed have attractive finite sample properties.

Indirect likelihood inference (revised)

Standard indirect Inference (II) estimators take a given finite-dimensional statistic, Z_{n} , and then estimate the parameters by matching the sample statistic with the model-implied population moment. We here propose a novel estimation method that utilizes all available information contained in the distribution of Z_{n} , not just its first moment. This is done by computing the likelihood of Z_{n}, and then estimating the parameters by either maximizing the likelihood or computing the posterior mean for a given prior of the parameters. These are referred to as the maximum indirect likelihood (MIL) and Bayesian Indirect Likelihood (BIL) estimators, respectively. We show that the IL estimators are first-order equivalent to the corresponding moment-based II estimator that employs the optimal weighting matrix. However, due to higher-order features of Z_{n} , the IL estimators are higher order efficient relative to the standard II estimator. The likelihood of Z_{n} will in general be unknown and so simulated versions of IL estimators are developed. Monte Carlo results for a structural auction model and a DSGE model show that the proposed estimators indeed have attractive finite sample properties.

Centile reference for total energy expenditure in infants from 1 to 12 months

Background: A knowledge of energy expenditure in infancy is required for the estimation of recommended daily amounts of food energy, for designing artificial infant feeds, and as a reference standard for studies of energy metabolism in disease states. Objectives: The objectives of this study were to construct centile reference charts for total energy expenditure (TEE) in infants across the first year of life. Methods: Repeated measures of TEE using the doubly labeled water technique were made in 162 infants at 1.5, 3, 6, 9 and 12 months. In total, 322 TEE measurements were obtained. The LMS method with maximum penalized likelihood was used to construct the centile reference charts. Centiles were constructed for TEE expressed as MJ/day and also expressed relative to body weight (BW) and fat-free mass (FFM). Results: TEE increased with age and was 1.40,1.86, 2.64, 3.07 and 3.65 MJ/day at 1.5, 3, 6, 9 and 12 months, respectively. The standard deviations were 0.43, 0.47, 0.52, 0.66 and 0.88, respectively. TEE in MJ/kg increased from 0.29 to 0.36 and in MJ/day/kg FFM from 0.36 to 0.48. Conclusions: We have presented centile reference charts for TEE expressed as MJ/day and expressed relative to BW and FFM in infants across the first year of life. There was a wide variation or biological scatter in TEE values seen at all ages. We suggest that these centile charts may be used to assess and possibly quantify abnormal energy metabolism in disease states in infants.

Local influence for Student-t partially linear models

In this paper we extend partial linear models with normal errors to Student-t errors Penalized likelihood equations are applied to derive the maximum likelihood estimates which appear to be robust against outlying observations in the sense of the Mahalanobis distance In order to study the sensitivity of the penalized estimates under some usual perturbation schemes in the model or data the local influence curvatures are derived and some diagnostic graphics are proposed A motivating example preliminary analyzed under normal errors is reanalyzed under Student-t errors The local influence approach is used to compare the sensitivity of the model estimates (C) 2010 Elsevier B V All rights reserved

Feasible estimation of generalized linear mixed models (GLMM) with weak dependency between groups

This paper presents a two-step pseudo likelihood estimation technique for generalized linear mixed models with the random effects being correlated between groups. The core idea is to deal with the intractable integrals in the likelihood function by multivariate Taylor's approximation. The accuracy of the estimation technique is assessed in a Monte-Carlo study. An application of it with a binary response variable is presented using a real data set on credit defaults from two Swedish banks. Thanks to the use of two-step estimation technique, the proposed algorithm outperforms conventional pseudo likelihood algorithms in terms of computational time.

Modelagem e estimação de parâmetros genéticos e fenotípicos para pesos do nascimento à seleção (378 dias) de machos Nelore

Genetic parameters for weights (17, 942 records), obtained in intervals of 60 days, from the birth to selection (378 days of age), of 2,582 males of the Nellore breed was estimated in univariate analyses by the Maximum Restricted Likelihood method. The models of analysis models included the fixed effects of contemporary groups, month of birth, mother age and age when the weights were collected as covariate. Three models differing in random effects were tested: the model 1 (M1) was adjusted for the direct and maternal addictive genetic effects and maternal permanent environment; in model 2 (M2) the maternal genetic effect was excluded; and the model 3 (M3) was only adjusted for the direct addictive genetic effect. The test of likelihood (LRT) detected significant differences, for all the ages, of M2 and M1 in relation to the simple model (M3), evidencing the importance of the maternal effects. Except for the birth weight (0.40), low values (0.05 to 0.12) of h(2) were found for M1 and M2 until 8 months of age and, after this period, reasonable increase could be observed, reaching 0.28 until 13 months of age. The estimates of the total variance fraction, due to the effect of maternal permanent environment, were high and practically became unaffected between the models 1 and 2. Maternal effects, not necessarily decomposed (in genetic addictive and permanent environment), affected the Nellore males growth. Models that contemplate maternal effects, besides the genetic addictive direct effects, are more realistic to describe the trajectory of the variances in the initial phases of growth of Nelore male calves.

Caracterização estatística de processos sísmicos via Distribuição Generalizada de Pareto. Estudo de caso: João Câmara-RN

The work is to make a brief discussion of methods to estimate the parameters of the Generalized Pareto distribution (GPD). Being addressed the following techniques: Moments (moments), Maximum Likelihood (MLE), Biased Probability Weighted Moments (PWMB), Unbiased Probability Weighted Moments (PWMU), Mean Power Density Divergence (MDPD), Median (MED), Pickands (PICKANDS), Maximum Penalized Likelihood (MPLE), Maximum Goodness-of-fit (MGF) and the Maximum Entropy (POME) technique, the focus of this manuscript. By way of illustration adjustments were made for the Generalized Pareto distribution, for a sequence of earthquakes intraplacas which occurred in the city of João Câmara in the northeastern region of Brazil, which was monitored continuously for two years (1987 and 1988). It was found that the MLE and POME were the most efficient methods, giving them basically mean squared errors. Based on the threshold of 1.5 degrees was estimated the seismic risk for the city, and estimated the level of return to earthquakes of intensity 1.5°, 2.0°, 2.5°, 3.0° and the most intense earthquake never registered in the city, which occurred in November 1986 with magnitude of about 5.2º

Evaluación genética de búfalos de la raza Murrah para la producción de leche en el día de control y a los 305 días de lactancia

There were analyzed 4757 complete lactations of the Murrah breed, daughters of 187 bulls, with the goal of verifying the viability upon employing the test-day (PDLC), on substitution of the milk yield at 305th day of lactation (PL305), in the genetic evaluations. The components of variance for the PDLC1 to PDLC9 and for the PL305 were estimated in uni-traits analysis according to maximum restricted likelihood method. The used model included the genetic direct additive random effects, of residual and permanent environment. There were considered as fixed effects, the contemporary group and the number of milkings and the age of the cow at the moment of parity co-variable (quadratic and linear effect). The contemporary groups were constituted by the herd-year-month of control for the PDLC and by herd-year-epoch of parity for PL305. The estimates of heritability for the PDLC and PL305 were 0.12 to 0.23 and 0.22, respectively. The correlations of order of the predicted genetic values for the 187 bulls, obtained between the PDLC and the PL305, were from moderate to high, varying from 67.74 to 83.12. From the minimum selection of the 10% of the best bulls relating to the predicted genetic value for the PL305, the coincidence among the classification of these animals was over 68%, when evaluated by the PDLC3,PDLC4,PDLC5 and PDLC6. Upon selecting the 5% of the best animals that coincidence presented a lower value.

Comparação entre metodologias para avaliar a idade à primeira concepção em éguas puro-sangue inglês

In Brazil, due to the breeding season for Thoroughbred, the reproductive data are normally truncate, since the breeders try to get animals that were born at the beginning of the breeding season in order to take their competitive advantages (more developed, mature and trained animals) compared to animals born later in the same breeding season. To analyze these data suitable methods should be used. Then, this paper aims to compare three methodologies: the method of maximum restricted likelihood, using MTDFREML, bayesian analysis without censured data by software MTGSAM and bayesian analysis with censured data by software LMCD, to evaluate age at first conception in thoroughbred mares, in order to verify its impact on the choice of stallions during selection. The database contained 3509 records for age at first conception (months) for thoroughbred mares. The heritability estimates were 0.23, 0.30 and 0.0926 (log scale), for MTDF, MTGSAM and LMCD, respectively. Considering all animals in the pedigree (6713), ranking correlations varied from 0.91 to 0.99. When only stallions were considered (656), those varied from 0.48 to 0.99 (considering different percentages of selected males) between evalua-tion methods. The highest changes in the general classification were observed when LMCD was compared to the other two methods. As the linear censured model is the most suitable for trait analysis with censured data, it was observed that censure information would lead to the choice of different animals during the selection process, when compared to the two other methodologies.

Interação genótipo-ambiente em características de crescimento de bovinos da raça Nelore

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Análise genética de características de crescimento e de carcaça em bovinos Nelore

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Avaliação genético-quantitativa de características reprodutivas em éguas da raça puro-sangue inglês

Pós-graduação em Zootecnia - FMVZ

Estimação de parâmetros genéticos para características produtivas em búfalos na Amazônia Oriental

Dados de 1.182 registros de produção de fêmeas bubalinas da raça Murrah e seus mestiços, parindo no período de 1967 a 2005, foram utilizados para estimação de parâmetros genéticos utilizando-se o método de máxima verossimilhança restrita. O modelo animal utilizado para estimação de componentes de variância incluiu os efeitos fixos de rebanho, ano e época de parto, ordem de parto e duração da lactação e os efeitos aleatórios do animal, e ambiente permanente e temporário. As estimativas de herdabilidade foram 0,25, 0,18, 0,08 e 0,09, para produção de leite, produção de gordura, duração da lactação e produção de leite por dia de intervalo de parto, respectivamente. As estimativas de repetibilidade foram 0,33, 0,29 e 0,10 para produção de leite, produção de gordura e duração da lactação, respectivamente. As correlações genéticas entre produções de leite e gordura, produção de leite com duração da lactação, produção de leite com produção de leite por dia de intervalo de partos, produção da gordura com duração da lactação, produção de gordura com produção de leite por dia de intervalo de partos e duração da lactação com produção de leite por dia de intervalo de partos foram 0,93; 0,76; 0,99; 0,89; 0,87 e -0,27, respectivamente. Os resultados demonstram que ganhos genéticos podem ser obtidos pela seleção das produções de leite e gordura.

GENETIC VARIABILITY FOR JUVENILE CHARACTERS OF Cedrela fissilis VELL. PROGENIES: A SUBSIDY FOR DEFINITION OF SEED COLLECTION ZONE

The objective of this study was to evaluate the genetic differences among three matrix groups of Cedrela fissilis based on quantitative juvenile variables on a progeny test to define seed collecting zones and use of seeds of this species in the study region as well as to evaluate genetic variability of the sampled material. A progeny test was established in a nursery with seeds from 48 seed trees collected in the municipalities of Rio Negrinho, Mafra and Sao Bento do Sul, state of Santa Catarina, and in the municipalities of Lapa, Rio Negro, Campo do Tenente and Antonio Olinto, state of Parana. Of the collected seed trees, 33 sampled trees were distributed in three sites and 15 trees were dispersed in the studied region. It was used a complete random block design, with 8 replicates and 20 plants per plot. Evaluated data included: emergency rate; seedling base diameter and height (61, 102 and 145 days after the seeds were sowed); seedling survival; number of leaves per seedling; aerial section dry mass and root dry mass; and the foliar area of the third fully expanded leaf measured from the apical meristem. The Maximum Restricted Likelihood Method (REML) was used, using the software SELEGEN for analysis. It was found that the juvenile characters are strongly genetically controlled and they can be used to estimate genetic variability of population samples of Cedrela fissilis. The three groups of trees spatially limited did not significantly differ among each other, allowing to conclude that the three areas are part of the same tree seed transfer zone.