Esta dissertação é formada dois ensaios com temas vinculados à educação. O primeiro ensaio investiga a evolução das desigualdades de oportunidades educacionais no Brasil sob o prisma das etnias e dos gêneros, com base na concepção de John Roemer (1998) de Igualdade de Oportunidades. Para quantificar a desigualdade são utilizados os índices de dissimilaridade de Yalonetzky (2010) aplicados aos microdados da PNAD de 2011. Para a amostra estudada, os resultados apontaram para a redução da desigualdade de oportunidades educacionais na dimensão étnica e para seu aumento, no âmbito dos gêneros, ao longo dos anos. O segundo ensaio busca avaliar o impacto da universidade pública sobre o seu egresso, medido principalmente sob a forma de rendimentos do trabalho. Através de modelos baseados na Equação de Mincer (1974), buscou-se verificar os principais determinantes da renda do egresso, utilizando-se, para tanto, uma amostra de dados de egressos da UFRGS. A análise apontou para conclusões que levam a crer que o diploma da universidade pública não é suficiente para, em termos de rendimento, compensar discriminações existentes no mercado de trabalho analisado. Outra contribuição do ensaio foi apontar, para algumas áreas do conhecimento, a remuneração média, a incidência de desemprego, a relevância atribuída pelo egresso à formação recebida e seu vínculo à atividade profissional exercida. / This dissertation is composed by two essays whose subjects are linked with education. The first one investigates evolution of the inequalities of opportunities of education in Brazil, through ethnical and genre prism, using the Opportunities Equality framework proposed by Roemer (1998). To quantify inequality, the dissimilarities indexes are used on PNAD 2011 data. The results pointed to a reduction of opportunity inequality in ethnical types, while it detected an increase in inequality for the gender types. The second essay aims to assess the impact of the public university in Brazil on its graduated, understood mainly as his wage. Using econometric specifications based on Mincer (1974), the study tried to find out the main determinants of graduated earnings, using a sample of graduated from UFRGS. The analysis suggested that the public university's diploma is not enough to compensate the discrimination existing in the studied labor market. Another contribution of this essay was to point, for some knowledge areas, the mean earnings, the unemployment rates, the relevance attributed by the graduated to the received education and its links with his professional career.

Made available in DSpace on 2014-08-20T13:59:09Z (GMT). No. of bitstreams: 1 dissertacao_leticia_benitez.pdf: 775594 bytes, checksum: 1d79ebc2ce3bb293397c9d3cbc98c0a0 (MD5) Previous issue date: 2008-07-17 / Plants under natural conditions are exposed to several environmental stresses which affect their metabolism. Among them, soil salinity is one of the most serious problems in irrigated agriculture. Trying to identify genetic variability for this character, the main goal of this work was to evaluate the germination and the initial development of plantlets of 10 rice genotypes under salt stress. The plantlets were cultivated ex vitro and in vitro, and many morphological characters were used to phenotype their responses. The investigations were performed ex vitro in the greenhouse and in vitro in culture media with NaCl concentrations of 0, 68, 136 and 204 mM added to a nutrient solution and to culture media, respectively. After 21 days from the beginning of each experiment, plantlet emergence and seed germination in vitro were evaluated. Other characters measured were shoot length, number of leaves, leaf area, root length, number of roots and fresh and dry mass of aerial part and radicular system. Analysis of variance, regression fitting, percentage of reduction and dissimilarity analysis were performed. The salinity damages were seen only at the initial phase of germination and plantlet emergence except on 204 mM concentration which was inhibitory to in vitro seed germination. All measured characters had their development reduced in saline substrate. The characters average shoot area, shoot biomass and root biomass were the most sensitive to NaCl and the characters leaf and root number were the less affected by the salinity excess. The genotypes showed different responses in both experimental conditions. The genotypes formed three groups when an UPGMA hierarchal method was applied to both system of cultivation. Five and two groups were found when Tocher optimization method was applied to in ex vitro and in vitro data, respectively. The variables average shoot area and average root dry mass contributed more to similarity between genotypes ex vitro, while the average shoot and root weight were the most contributing for the in vitro dissimilarity. The variable shoot length was the one that less contributed to genotype discrimination in both cultivation systems. It is concluded that the genotype BRS Bojuru presented the highest tolerance to salinity while BRS Ligeirinho showed the lowest tolerance in both investigations. / As plantas, sob condições naturais, estão expostas a vários estresses ambientais que afetam seu metabolismo. Dentre estes, a salinidade dos solos e da água de irrigação é um dos mais sérios problemas para a agricultura irrigada. Sabendo que o germoplasma do arroz possui uma variabilidade genética para tolerância à salinidade, o objetivo deste trabalho, foi avaliar a germinação e o desenvolvimento inicial de plântulas de 10 genótipos de arroz, cultivados ex vitro e in vitro, por meio de caracteres morfológicos e agrupá-los para o caráter tolerância à salinidade. Foram realizados trabalhos em casa de vegetação e no sistema de cultura in vitro com as concentrações de 0, 68, 136 e 204 mM de NaCl acrescidos à solução nutritiva e ao meio de cultura, respectivamente. Após 21 dias do início de cada experimento, foram avaliadas a emergência de plântulas ex vitro e a germinação de sementes in vitro, além das médias dos caracteres altura da parte aérea, número de folhas, área foliar, comprimento de raiz, número de raiz e massa fresca e seca da parte aérea e do sistema radicular. Foram procedidas análises de variância, ajuste de regressões, cálculos de percentagem de redução e análise de dissimilaridade. A salinidade a atrasou a fase inicial de germinação e emergência de plântulas, exceto na concentração de 204 mM que foi inibitória para germinação das sementes in vitro. Todos os caracteres mensurados tiveram seu desenvolvimento reduzido em substrato salino, sendo os caracteres correspondentes à área foliar média e à fitomassa média da parte aérea e do sistema radicular os mais suscetíveis ao NaCl e número médio de folhas e de raiz os menos afetados pelo excesso de salinidade. Observou-se dissimilaridade entre os genótipos estudados para tolerância à salinidade nas duas condições experimentais, verificada pela formação de três grupos pelo método hierárquico UPGMA, em ambos sistemas de cultivo, e de cinco e dois grupos pelo método de otimização de Tocher no cultivo ex vitro e in vitro, respectivamente. As variáveis, área foliar média e massa seca média de raiz, contribuíram mais para a dissimilaridade entre os genótipos ex vitro, enquanto que massa fresca média da parte aérea e de raiz foram as que mais contribuíram para a dissimilaridade in vitro. A variável altura média da parte aérea foi a que menos contribuiu para a separação dos genótipos nos dois sistemas de cultivo. Pode-se concluir que o genótipo BRS Bojuru apresentou maior tolerância à salinidade, enquanto que BRS Ligeirinho mostrou maior sensibilidade nos dois trabalhos.

The Gray Fossil Site, northeastern Tennessee, is formed by multiple karst sub-basins filled with lacustrine sediments. The oldest sediments found were recently dated as Paleo-Eocene by palynological means, whereas the youngest sediments are considered Mio-Pliocene based on their faunal assemblage. In this study, we examined the Mio-Pliocene lacustrine sediments from the Gray Fossil Site to determine the Late Neogene floral characteristics of a site within the southern Appalachian Mountains. The Mio-Pliocene lacustrine sinkhole fill preserves a unique fossil assemblage, which includes invertebrate, vertebrate, and floral remains. Floral remains are represented by wood, seeds, leaves, and pollen grains. Forty-seven palynological samples from six different test-pits were analyzed. All pits exhibit a low pollen yield, a result of basic pH levels, drought, and fire events that occurred during deposition. The palynofloral assemblage has a low to moderate diversity and is largely dominated by a Quercus-Carya-Pinus assemblage (~ 90% of the palynoflora). The presence of Pterocarya grains supports a Late Neogene age for these lacustrine sediments. Comparison with modern pollen-based floras from North America suggests that: (1) examined pits can be discriminated into two separate groups based on their palynofloral signatures, (2) the Mio-Pliocene vegetation at the site ranged between a closed to open woodland setting, depending on the intensity and frequency of drought and fire events, and (3) the fossil palynofloral assemblage is comparable to what would be expected in the modern North American Mesophytic Forest region.

When analyzing ecological data, one considers traditional multivariate techniques to be unsuitable. The use of dissimilarity coefficients and distance matrices is a way, how to solve this problem. In this work we present some of these coefficients and distance-based tests: Mantel test, several versions of ANOSIM and MRPP tests and distance-based test for homogeneity of multivariate dispersions. We focus on relationships among these tests and illustrate the use with an example. We also discuss the difficulties of interpretation of the results of these tests.

Made available in DSpace on 2016-03-15T19:25:50Z (GMT). No. of bitstreams: 1 Nereida Silveria.pdf: 1043135 bytes, checksum: 77740a92d7c758cec23075ec4c0f5ecb (MD5) Previous issue date: 2007-02-06 / Fundo Mackenzie de Pesquisa / The study investigated the effect of similarity and dissimilarity in characteristics of different depth levels, perceptual or real, over the superiors and subordinate relationship quality through LMX (Leader-Member Exchange) viewpoint. The characteristics investigated were gender and age at the superficial level and work family conflict at deep level. The study counted with the comparative data of 89 dyads. The results indicate the influence of general similarity perception in the quality of superiors and subordinate relationship. The stronger the similarity perceived, better the dyadic relationship. This effect increases when subordinate declares satisfied with the quality and frequency of the contact with his/her superior. Results demonstrate that similarity perception does not correspond to real similarity or dissimilarity, but this correspondence is enhanced when the contact is satisfactory. / O estudo buscou investigar os efeitos da similaridade e dissimilaridade em características de diferentes níveis de profundidade, reais ou percebidas, na qualidade da relação de superiores(as) e subordinados(as) sob a ótica da LMX (Leader-Member Exchange). As características investigadas foram gênero e idade, no nível superficial e conflito família-trabalho no nível profundo. O estudo contou com os dados comparativos de 89 díades. Os dados indicam a influência da percepção de similaridade geral na qualidade da relação de superiores e subordinados. Quanto maior a similaridade percebida, melhor a relação diádica. Este efeito é ampliado quando o(a) subordinado(a) se declara satisfeito(a) com a qualidade e a freqüência do contato com seu/sua superior(a). Resultados demonstram que a percepção de similaridade não corresponde à similaridade ou dissimilaridade real, mas esta correspondência se amplia quando o contato for satisfatório.

The need to quantify distance between two groups of objects is prevalent throughout the signal processing world. The difference of group means computed using the Euclidean, or L2 distance, is one of the predominant distance measures used to compare feature vectors and groups of vectors, but many problems arise with it when high data dimensionality is present. Maximum mean discrepancy (MMD) is a recent unsupervised kernel-based pattern recognition method which may improve differentiation between two distinct populations over many commonly used methods such as the difference of means, when paired with the proper feature representations and kernels. MMD-based distance computation combines many powerful concepts from the machine learning literature, such as data distribution-leveraging similarity measures and kernel methods for machine learning. Due to this heritage, we posit that dissimilarity-based classification and changepoint detection using MMD can lead to enhanced separation between different populations. To test this hypothesis, we conduct studies comparing MMD and the difference of means in two subareas of image analysis and understanding: first, to detect scene changes in video in an unsupervised manner, and secondly, in the biomedical imaging field, using clinical ultrasound to assess tumor response to treatment. We leverage effective computer vision data descriptors, such as the bag-of-visual-words and sparse combinations of SIFT descriptors, and choose from an assessment of several similarity kernels (e.g. Histogram Intersection, Radial Basis Function) in order to engineer useful systems using MMD. Promising improvements over the difference of means, measured primarily using precision/recall for scene change detection, and k-nearest neighbour classification accuracy for tumor response assessment, are obtained in both applications.

We consider the problem of learning the parameters of a structured output prediction model, that is, learning to predict elements of a complex interdependent output space that correspond to a given input. Unlike many of the existing approaches, we focus on the weakly supervised setting, where most (or all) of the training samples have only been partially annotated. Given such a weakly supervised dataset, our goal is to estimate accurate parameters of the model by minimizing the regularized empirical risk, where the risk is measured by a user-specified loss function. This task has previously been addressed by the well-known latent support vector machine (latent SVM) framework. We argue that, while latent SVM offers a computational efficient solution to loss-based weakly supervised learning, it suffers from the following three drawbacks: (i) the optimization problem corresponding to latent SVM is a difference-of-convex program, which is non-convex, and hence susceptible to bad local minimum solutions; (ii) the prediction rule of latent SVM only relies on the most likely value of the latent variables, and not the uncertainty in the latent variable values; and (iii) the loss function used to measure the risk is restricted to be independent of true (unknown) value of the latent variables. We address the the aforementioned drawbacks using three novel contributions. First, inspired by human learning, we design an automatic self-paced learning algorithm for latent SVM, which builds on the intuition that the learner should be presented in the training samples in a meaningful order that facilitates learning: starting frome easy samples and gradually moving to harder samples. Our algorithm simultaneously selects the easy samples and updates the parameters at each iteration by solving a biconvex optimization problem. Second, we propose a new family of LVMs called max-margin min-entropy (M3E) models, which includes latent SVM as a special case. Given an input, an M3E model predicts the output with the smallest corresponding Renyi entropy of generalized distribution, which relies not only on the probability of the output but also the uncertainty of the latent variable values. Third, we propose a novel learning framework for learning with general loss functions that may depend on the latent variables. Specifically, our framework simultaneously estimates two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. We demonstrate the efficacy of our contributions on standard machine learning applications using publicly available datasets.

Le projet SoC-Trace a pour objectif le développement d'un ensemble de méthodes et d'outils basés sur les traces d'éxécution d'applications embarquées multicoeur afin de répondre aux besoins croissants d'observabilité et de 'débogabilité' requis par l'industrie. Le projet vise en particulier le développement de nouvelles méthodes d'analyse, s'appuyant sur différentes techniques d'analyse de données telles que l'analyse probabiliste, la fouille de données, et l'agrégation de données. Elles devraient permettre l'identification automatique d'anomalies,l'analyse des corrélations et dépendances complexes entre plusieurs composants d'une application embarquées ainsi que la maîtrise du volume important des traces qui peut désormais dépasser le GigaOctet. L'objectif de la thèse est de fournir une représentation de haut niveau des informations contenues dans les traces, basée sur la sémantique. Il s'agira dans un premier temps de développer un outil efficace de comparaison entre traces;de définir une distance démantique adaptée aux traces, puis dans un second temps d'analyser et d'interpréter les résultats des comparaisons de traces en se basant sur la distance définie. / The SoC-Trace project aims to develop a set of methods and tools based on execution traces of multicore embedded applications to meet the growing needs of observability and 'débogability' required by the industry. The project aims in particular the development of new analytical methods, based on different data analysis techniques such as probabilistic analysis, data mining, and data aggregation. They should allow the automatic identification of anomalies, the analysis of complex correlations and dependencies between different components of an embedded application and control of the volume traces that can now exceed the gigabyte. The aim of the thesis is to provide a high-level representation of information in the trace based semantics. It will initially develop an effective tool for comparing traces, to define a semantic distance for execution traces, then a second time to analyze and interpret the results of comparisons of traces based on the defined distance.

The social inequalities in the Brazilians cities, as historical products express themselves in a heterogeneous way on territory from an arrangement of socioeconomic, structural and environmental factors, which define its quality of life, and they are related to the way as the concentration of the capital felt in the society starting from the adoption of the capitalist model. To understand the aspect of the dissimilarity in the cities, this research proposes to approach the spatial distribution from the urban infrastructure of basic sanitation (water distribution, waste management and sanitary sewer) to infer the existence of an association between the infrastructure presence and the income, and between these ones with population s density. In order to broaden discussion, the work also explores the political, economical, ambiental and social dimensions of infrastructure in the cities, comprehended as a sum between works and actions of State s responsibilities and as a result of public policy that have to redistribute service delivery an attend the values of social equality and democracy. According Habitat Agenda the planning systems have to integrate government actions to promote infrastructure services on the human settlements. Although, the equal urban infrastructure s distribution in Brazilian cities, in its current precarious condition, is one of the public administration s challenge, being then necessary the knowledge of the reality and the election of priorities in agreement with the beginning of the redistributive politics. With this fundamental, the research was applied in the urban portion of the municipal district of Maceió, capital of the state of Alagoas, area recognized as unequal in income and in infrastructure distribution; and it consisted of using, as methodological tool, the systematic and spatially analysis, in context of the techniques of geoprocessing and statistics data treatment, for measure, locate and to check the existents dissimilarity and also observe how the unequal distribution of urban infrastructure affects with a higher grade the low-incoming population. For this analysis were used data census of the year of 2000, which refer to domiciliary infrastructure and population data, organized by subjects and associated by a Geographic Information System (GIS). By utilizing a multi-criteria analysis the methodological process generate an indicative mapping showing the priority areas for public investments. Among the conclusions obtained by the accomplished analysis consists the confirmation of the existent inequality in the distribution of the income, of the population on territory and of the infrastructure in the city of Maceió, results of the historical and social development, and mainly to confirm of the relationship between the presence of the infrastructure and the income of the population. / As desigualdades sociais existentes nas cidades brasileiras, compreendidas como resultado de seu processo histórico, manifestam-se no território de forma heterogênea a partir de uma composição de fatores sócio-econômicos, estruturais e ambientais que definem o nível de qualidade de vida urbana, e estão relacionadas ao modo como a concentração do capital se deu na sociedade a partir da adoção do modelo capitalista. Para compreender como acontecem as desigualdades existentes nas cidades, esta dissertação se propôs abordar o aspecto da distribuição da infra-estrutura de saneamento básico (fornecimento de água, coleta de lixo e esgotamento sanitário) e suas associações com a renda da população e a densidade populacional. Inicialmente foram exploradas as dimensões política, econômica, ambiental e social da infra-estrutura da cidade, entendidas como um conjunto de obras e ações de responsabilidade do Estado, resultado de políticas públicas que devem buscar a melhoria da qualidade de vida nas cidades, de forma igualitária e democrática. Para isso, concordando com a Agenda Habitat, os sistemas de planejamento devem integrar ações governamentais para a promoção dos serviços e da infra-estrutura nos assentamentos humanos. No entanto, a distribuição eqüitativa da infra-estrutura urbana nas cidades brasileiras, no seu atual estado de precariedade, é um dos desafios para as gestões públicas, sendo então necessário o conhecimento da realidade e a eleição de prioridades de acordo com o princípio da política redistributiva. Com este fundamento, a pesquisa foi aplicada na porção urbana do município de Maceió, capital do estado de Alagoas, região reconhecida como desigual em renda e em distribuição de infra-estrutura; e consistiu em utilizar, como ferramenta metodológica a análise espacial sistêmica, através do geoprocessamento, e a análise estatística, com as quais foi possível medir, aferir e localizar as desigualdades existentes, e observar o quanto a distribuição desigual da infra-estrutura urbana afeta a população de menor renda e atende à demanda populacional. Para tal foram trabalhados dados do Censo do ano de 2000 referentes à infra-estrutura dos domicílios e dados populacionais, todos tratados estatisticamente por temas, tendo como unidade territorial o setor censitário. Através de Sistemas de Informações Geográficas (SIGs) foram gerados mapas temáticos, e utilizando a álgebra de mapas, foi realizada uma avaliação multicritério ou multiclassificatória a partir das três principais temáticas trabalhadas, produzindo o mapa temático das áreas prioritárias para investimentos em infra-estrutura. Este último produto foi realizado com o intuito de apontar os diversos níveis de carências existentes na cidade de forma a atender os princípios da equidade social, servindo como subsídio para a distribuição igualitária e democrática dos serviços urbanos de infra-estrutura. Dentre as conclusões obtidas pela análise realizada consta a confirmação da desigualdade existente na distribuição da renda, da população no território e da infra-estrutura na cidade de Maceió, fruto de seu desenvolvimento histórico e social, e principalmente a confirmação da relação entre a presença da infra-estrutura e a renda da população.

Submitted by Lara Gomes Melchior null (lahtuti@hotmail.com) on 2016-04-14T21:11:04Z No. of bitstreams: 1 Mestrado Lara Versão Final.pdf: 949503 bytes, checksum: 2675776db95f4fd4b4d6542497c65b45 (MD5) / Approved for entry into archive by Juliano Benedito Ferreira (julianoferreira@reitoria.unesp.br) on 2016-04-15T13:12:06Z (GMT) No. of bitstreams: 1 melchior_lg_me_sjrp.pdf: 949503 bytes, checksum: 2675776db95f4fd4b4d6542497c65b45 (MD5) / Made available in DSpace on 2016-04-15T13:12:06Z (GMT). No. of bitstreams: 1 melchior_lg_me_sjrp.pdf: 949503 bytes, checksum: 2675776db95f4fd4b4d6542497c65b45 (MD5) Previous issue date: 2016-03-23 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / A escolha da unidade amostral e extensão da área de estudo é crucial para a compreensão dos processos e mecanismos influenciando os padrões de distribuição da biodiversidade. Dependendo da escala espacial escolhida, os padrões de distribuição serão influenciados por processos ecológicos e biogeográficos diferentes. Neste trabalho, nós alternamos as unidades amostrais e extensões da área estudada (corpos d’água em um município, corpos d’água em uma ecoregião, municípios em uma ecoregião e municípios na área total de estudo) para avaliar a importância relativa das variáveis ecológicas e biogeográficas na diversidade beta de anuros na Floresta Atlântica Brasileira. Para isso, nós parcionamos a diversidade beta total nos componentes substituição espacial de espécies e aninhamento. Nós observamos que, mantendo a mesma unidade amostral e aumentando a extensão da área de estudo, os valores da substituição espécies aumentam. Por outro lado, mantendo a mesma extensão e aumentando a unidade amostral, os valores da substituição de espécies diminuem. Além disso, nossos resultados concordam que a importância relativa dos fatores ecológicos e biogeográficos na diversidade beta de anuros é dependente da escala espacial. Considerando as menores unidade amostral e extensão, as variáveis ambientais e distância geográfica se alternam na importância relativa associada aos valores de diversidade beta, enquanto que, em escalas espaciais maiores, as ecoregiões passam a ter uma importância relativa maior. / The selection of the sampling unit and extension of the study area is crucial for understanding the processes and mechanisms influencing biodiversity distribution patterns. Depending on the chosen scale, patterns of distribution will be influenced by different ecological and biogegraphical processes. Here, we alternate the sample units and extensions of the study area (water bodies in a municipality, water bodies in an ecoregion, municipalities in an ecoregion, and municipalities in total area of study) to assess the relative importance of ecological and biogeographical factors in the beta diversity of frogs in the Brazilian Atlantic Forest. We partitioned total beta diversity into two additive components, spatial species turnover and nestedness. We found that considering the same sample unit and increasing the extent of the study area the turnover values increased. On the other hand, considering the same extension and increasing the sampling unit, the turnover values decreased. Furthermore, our results agree that relative importance of the ecological and biogeographical processes in frog beta diversity is dependent on spatial scales. Considering the smallest sampling unit and extension, environmental variables and geographic distance alternated in their relative importance associated with beta diversity, while at largest spatial scales, ecoregions showed higher values of relative importance.