Plusieurs auteurs ont proposé récemment des modèles et des algorithmes pour l'estimation nonparamétrique de mélanges multivariés finis dont l'identifiabilité n'est pas toujours assurée. Entre les modèles considérés, l'hypothèse des coordonnées indépendantes conditionnelles à la sous-population de provenance des individus fait l'objet d'une attention croissante, en raison des développements théoriques et pratiques envisageables, particulièrement avec la multiplicité des variables qui entrent en jeu dans le framework statistique moderne. Dans ce travail, nous considérons d'abord un modèle plus général supposant l'indépendance, conditionnellement à la composante, de blocs multivariés de coordonnées au lieu de coordonnées univariées, permettant toute structure de dépendance à l'intérieur de ces blocs. Par conséquent, les fonctions de densité des blocs sont complètement multivariées et non paramétriques. Nous présentons des arguments d'identifiabilité et introduisons pour l'estimation dans ce modèle deux algorithmes méthodologiques dont les procédures de calcul ressemblent à un véritable algorithme EM mais incluent une étape additionnelle d'estimation de densité: un algorithme rapide montrant l'efficacité empirique sans justification théorique, et un algorithme lissé possédant une propriété de monotonie comme certain algorithme EM, mais plus exigeant en terme de calcul. Nous discutons également les méthodes efficaces en temps de calcul pour l'estimation et proposons quelques stratégies. Ensuite, nous considérons une extension multivariée des modèles de mélange utilisés dans le cadre de tests d'hypothèses multiples, permettant une nouvelle version multivariée de contrôle du False Discovery Rate. Nous proposons une version contrainte de notre algorithme précédent, adaptée spécialement à ce modèle. Le comportement des algorithmes de type EM que nous proposons est étudié numériquement dans plusieurs expérimentations de Monte Carlo et sur des données réelles de grande dimension et comparé avec les méthodes existantes dans la littérature. En n, les codes de nos nouveaux algorithmes sont progressivement ajoutés sous forme de nouvelles fonctions dans le package en libre accès mixtools pour le logiciel de statistique R. / Recently several authors have proposed models and estimation algorithms for finite nonparametric multivariate mixtures, whose identifiability is typically not obvious. Among the considered models, the assumption of independent coordinates conditional on the subpopulation from which each observation is drawn is subject of an increasing attention, in view of the theoretical and practical developments it allows, particularly with multiplicity of variables coming into play in the modern statistical framework. In this work we first consider a more general model assuming independence, conditional on the component, of multivariate blocks of coordinates instead of univariate coordinates, allowing for any dependence structure within these blocks. Consequently, the density functions of these blocks are completely multivariate and nonparametric. We present identifiability arguments and introduce for estimation in this model two methodological algorithms whose computational procedures resemble a true EM algorithm but include an additional density estimation step: a fast algorithm showing empirical efficiency without theoretical justification, and a smoothed algorithm possessing a monotony property as any EM algorithm does, but more computationally demanding. We also discuss computationally efficient methods for estimation and derive some strategies. Next, we consider a multivariate extension of the mixture models used in the framework of multiple hypothesis testings, allowing for a new multivariate version of the False Discovery Rate control. We propose a constrained version of our previous algorithm, specifically designed for this model. The behavior of the EM-type algorithms we propose is studied numerically through several Monte Carlo experiments and high dimensional real data, and compared with existing methods in the literature. Finally, the codes of our new algorithms are progressively implemented as new functions in the publicly-available package mixtools for the R statistical software.

Kernel Methods in Particle Filtering David Coufal Doctoral thesis - abstract The thesis deals with the use of kernel density estimates in particle filtering. In particular, it examines the convergence of the kernel density estimates to the filtering densities. The estimates are constructed on the basis of an out- put from particle filtering. It is proved theoretically that using the standard kernel density estimation methodology is effective in the context of particle filtering, although particle filtering does not produce random samples from the filtering densities. The main theoretical results are: 1) specification of the upper bounds on the MISE error of the estimates of the filtering densities and their partial derivatives; 2) specification of the related lower bounds and 3) providing a suitable tool for checking persistence of the Sobolev character of the filtering densities over time. In addition, the thesis also focuses on designing kernels suitable for practical use. 1

La miniaturisation technologique augmente la sensibilité des circuits intégrés auxdéfauts et nous observons à chaque nouvelle génération technologique une dégradation rapidedu rendement de fabrication et de la fiabilité. Les mémoires occupent la plus grande partie dela surface des SoCs et contiennent la vaste majorité des transistors. De plus, pour augmenterleur densité elles sont conçues de façon très serrée. Elles concentrent ainsi la plus grandepartie des défauts de fabrication et représentent aussi les parties les plus sensibles face auxperturbations. Elles sont par conséquent les parties des SoCs les plus affectées par ladégradation du rendement de fabrication et de la fiabilité. L’objectif de cette thèse est deproposer des architectures combinant de façon optimale : algorithmes de test, architecturesBIST, et codes correcteurs d’erreurs afin de proposer des solutions efficaces pourl’amélioration du rendement de fabrication et de la fiabilité des mémoires embarquées. / Nanometric scaling increases the sensitivity of integrated circuits to defects andperturbations. Thus, each new generation of manufacturing process is accompanied by a rapiddegradation of manufacturing yield and reliability. Embedded memories occupy the largestpart of the area of SoCs and comprise the vast majority of transistors. In addition, forincreasing the integration density, they are designed very tightly to the design and electricalrules. Hence, embedded memories concentrate the majority of the manufacturing defectsaffecting a SoC, and are also more sensitive to perturbations. Thus, they are the parts of theSoC the most affected by the deterioration of manufacturing yield and reliability. This thesisdevelops repair architectures optimally combining test algorithms, BIST architectures, anderror correcting codes, in order to propose effective solutions for improving themanufacturing yield and reliability of embedded memories affected by high defect densities.

Made available in DSpace on 2014-06-11T19:22:23Z (GMT). No. of bitstreams: 0 Previous issue date: 2006-06-05Bitstream added on 2014-06-13T18:08:11Z : No. of bitstreams: 1 rodrigues_vcs_me_jabo.pdf: 201320 bytes, checksum: 0b3a492d89a4de0571bff67b2f0754ed (MD5) / Com o intuito de conhecer o potencial zootécnico do pampo, foi avaliado o efeito de diferentes densidades de estocagem (100 e 200 peixes/m3) na sobrevivência e crescimento do pampo amarelo (Trachinotus carolinus), criado em tanque-rede instalado em ambiente marinho. Peixes com aproximadamente 1,0 a 2,0 g, capturados em arrastos nas praias rasas do município de Ubatuba, SP, foram transportados ao Núcleo de Pesquisa e Desenvolvimento do Litoral Norte - Agência Paulista Tecnológica do Agronegócio - Secretaria da Agricultura e Abastecimento e mantidos durante 45 dias em tanques-rede de 5mm de malhagem. Após seleção manual, animais de 4,0 ± 0,17g foram estocados, nas densidades de 100 e 200 peixes/m3, em tanques-rede de multifilamento com 2 x 2 x 1,5m e 12 mm de malhagem, fixos a um deck flutuante (6,0 x 8,0 m), instalado diretamente no mar a profundidade de aproximadamente 3,5 ± 0,5 m. Os animais foram alimentados duas vezes ao dia até a saciação com ração de 40% de PB. Após 120 dias, as taxas de sobrevivência média observadas foram 77,23% e 73,11 %; pesos médios (52,79 ± 4.86 9 e 52,64 ± 5.64 g); Ganho de Peso (GP) (48,69 ± 4,05 9 e 48,73 ± 4,00 g); Taxa de Crescimento Especifico (TCE) (2,13 ± 1,24 e 2,16 ± 1,03) e Conversão Alimentar Aparente (CAA) (3,23 ± 1,52 e 2,66 ± 0,78), respectivamente para 100 e 200 peixes/m3. Estes dados não apresentaram diferenças significativas (P~0,05) entre os tratamentos testados. Entretanto os ganhos de peso do pampo, no presente experimento foram superiores a muitos trabalhos encontrados na literatura, o que nos leva a concluir que a densidade de 200 peixes/m3, utilizada neste estudo, pode ainda estar abaixo da capacidade de suporte para a espécie e pode ser o ponto de partida para futuros estudos, considerando o estágio de desenvolvimento dos animais. / Aiming to know the zoothecnic potential of pampo, it was evaluated the effect of stocking densities (100 and 200 fish/m3) in the survival and growth rates of Florida pampano (Trachinotus carolinus), reared in marine cage. Fish with approximately 1 - 2g captured by seine in flat beaches of the city of Ubatuba, S.P. were transferred to the Núcleo de Pesquisa e Desenvolvimento do Litoral Norte - Agencia Paulista Tecnologica do Agronegocio - Secretaria Agricultura e Abastecimento and kept during 45 days in cage of 5mm mesh size. After manual selection, animals of 4.0 ± 0.17g had been stored, in the studied densities, cage of multifilament with 2 x 2 x 1.5 m and 12 mm of mesh size fixed to adeck (6.0 8.0 m), installed directly in the sea at a depth of approximately 3.5 ± 0.5 m. The animals had been fed twice¬-a day until satiation with ration of 40% of PB. Soon after a 120 days experimental period, the average survival rate (77.23% and 73.11 %); average body-weight (52.79 ± 4.86 and 52.64 ± 5.64); weiht gain (48.69 ± 4.05 and 48.73 ± 4.00); daily growth increment (2.13 ± 1.24 and 2.16±1.03) and apparent feed conversion rate (3.23 ± 1.52 and 2.66 ± 0.78), respectively for 100 and 200 fish/m3,were recorded. These parameters didn't present significant differences (P>=0,05) between the tested treatments. However pampo weight gain, in this experiment have been superior to many works found in literature, what allows the conclusion that pampo can be reaise in sea cages in densities above 200 fish/m3, and this density could be the starting point for future studies.

Orientador: João Antonio da Costa Andrade / Banca: Pedro Cesar dos Santos / Banca: José Branco de Miranda Filho / Resumo: Técnicas como a redução do espaçamento entre linhas, permitindo um melhor arranjo das plantas no campo, juntamente com o aumento da densidade de semeadura, podem ser empregadas para aumentar a interceptação da radiação solar, visando o aumento do rendimento para algumas cultivares de milho. O objetivo deste trabalho foi quantificar a variação genética do composto Flintisa Anão de milho, a fim de verificar se o mesmo apresenta potencial para melhoramento em condições de safrinha, espaçamento reduzido e duas densidades de semeadura. Foram avaliadas, na safrinha/2004, 150 progênies de meios irmãos, no espaçamento de 0,45 m entre linhas e nas densidades de 57.800 e 80.000 plantas ha-1, para os caracteres florescimento feminino, altura de plantas, altura de espigas, tombamento, prolificidade, grãos ardidos e rendimento de grãos. Variância genética aditiva, herdabilidade, correlação entre os caracteres, resposta correlacionada de um caráter mediante seleção em outro e ganho esperado com seleção, foram estimados para cada densidade e conjuntamente. Foram estimados ganhos de 16,0 e 19,2% para rendimento, 11,1 e 10,5% para prolificidade e 12,3 e 12,9% para altura de espigas, respectivamente para as densidades de 57.800 e 80.000 plantas ha-1. As estimativas de herdabilidade para altura de plantas, altura de espigas e rendimento, nas densidades 57.800 e 80.000 plantas ha-1 foram respectivamente 65,19 e 61,27%, 64,30 e 66,86% e 53,53 e 63,32%. As estimativas parecidas, e a ausência de interação progênies x densidades, indicam que não há necessidade de programas de seleção distintos. A densidade de 57.800 plantas ha-1 pode ser preferida apenas pelo fato das progênies...(Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Techniques as spacing row reduction, allowing a better plant arrangement in the field, together with the increase of the sowing density, can be used to increase the interception of the solar radiation, seeking the increase of the grain yield for some maize cultivars. The objective of this work was to quantify the genetic variation of the Dwarf Flintisa composite of maize, to verify its potential for improvement in off season crop, in reduced spacing between rows and two sowing densities. In the off season crop (March to July/2004) 150 half sib progenies were evaluated, in 0.45 m spacing rows and densities of 57800 and 80000 plants ha-1, for the traits female flowering, plant height, ear height, lodging, prolificacy, burning grains and grain yield. Additive genetic variance, heritability, correlation between traits, correlated response of a trait by selection in other and expected gain from selection, were estimated for each density and jointly. Gains of 16.0 and 19.2% for grain yield, 11.1 and 10.5% for prolificacy and 12.3 and 12.9% for ear height, respectively for the populations of 57800 and 80000 plants ha-1 were estimated. The heritability estimates for plant height, ear height and grain yield, in densities 57800 and 80000 plants ha-1 were 65.19 and 61.27%, 64.30 and 66.86% and 53.53 and 63.32%, respectively. The similar estimates and the absence of progenies x densities interaction, indicate that there different selection programs are not justified. The density of 57800 plants ha-1 can just be preferred because the progenies present larger grain yield, smaller lodging and higher prolificacy, on average. Therefore the Dwarf Flintisa composite has sufficient genetic variability for improvement for low and for high sowing density, in off season crop conditions...(Complete abstract click electronic access below) / Mestre

Made available in DSpace on 2014-06-11T19:29:44Z (GMT). No. of bitstreams: 0 Previous issue date: 2005-07-22Bitstream added on 2014-06-13T19:18:10Z : No. of bitstreams: 1 candido_ls_me_ilha.pdf: 332984 bytes, checksum: 467381227e709e8b87b6b03749be5529 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Técnicas como a redução do espaçamento entre linhas, permitindo um melhor arranjo das plantas no campo, juntamente com o aumento da densidade de semeadura, podem ser empregadas para aumentar a interceptação da radiação solar, visando o aumento do rendimento para algumas cultivares de milho. O objetivo deste trabalho foi quantificar a variação genética do composto Flintisa Anão de milho, a fim de verificar se o mesmo apresenta potencial para melhoramento em condições de safrinha, espaçamento reduzido e duas densidades de semeadura. Foram avaliadas, na safrinha/2004, 150 progênies de meios irmãos, no espaçamento de 0,45 m entre linhas e nas densidades de 57.800 e 80.000 plantas ha-1, para os caracteres florescimento feminino, altura de plantas, altura de espigas, tombamento, prolificidade, grãos ardidos e rendimento de grãos. Variância genética aditiva, herdabilidade, correlação entre os caracteres, resposta correlacionada de um caráter mediante seleção em outro e ganho esperado com seleção, foram estimados para cada densidade e conjuntamente. Foram estimados ganhos de 16,0 e 19,2% para rendimento, 11,1 e 10,5% para prolificidade e 12,3 e 12,9% para altura de espigas, respectivamente para as densidades de 57.800 e 80.000 plantas ha-1. As estimativas de herdabilidade para altura de plantas, altura de espigas e rendimento, nas densidades 57.800 e 80.000 plantas ha-1 foram respectivamente 65,19 e 61,27%, 64,30 e 66,86% e 53,53 e 63,32%. As estimativas parecidas, e a ausência de interação progênies x densidades, indicam que não há necessidade de programas de seleção distintos. A densidade de 57.800 plantas ha-1 pode ser preferida apenas pelo fato das progênies... / Techniques as spacing row reduction, allowing a better plant arrangement in the field, together with the increase of the sowing density, can be used to increase the interception of the solar radiation, seeking the increase of the grain yield for some maize cultivars. The objective of this work was to quantify the genetic variation of the Dwarf Flintisa composite of maize, to verify its potential for improvement in off season crop, in reduced spacing between rows and two sowing densities. In the off season crop (March to July/2004) 150 half sib progenies were evaluated, in 0.45 m spacing rows and densities of 57800 and 80000 plants ha-1, for the traits female flowering, plant height, ear height, lodging, prolificacy, burning grains and grain yield. Additive genetic variance, heritability, correlation between traits, correlated response of a trait by selection in other and expected gain from selection, were estimated for each density and jointly. Gains of 16.0 and 19.2% for grain yield, 11.1 and 10.5% for prolificacy and 12.3 and 12.9% for ear height, respectively for the populations of 57800 and 80000 plants ha-1 were estimated. The heritability estimates for plant height, ear height and grain yield, in densities 57800 and 80000 plants ha-1 were 65.19 and 61.27%, 64.30 and 66.86% and 53.53 and 63.32%, respectively. The similar estimates and the absence of progenies x densities interaction, indicate that there different selection programs are not justified. The density of 57800 plants ha-1 can just be preferred because the progenies present larger grain yield, smaller lodging and higher prolificacy, on average. Therefore the Dwarf Flintisa composite has sufficient genetic variability for improvement for low and for high sowing density, in off season crop conditions...(Complete abstract click electronic access below)

Com o objetivo de estudar o comportamento fenológico, morfológico e produtivo de cultivares de algodoeiro com diferentes arquiteturas de plantas em diferentes espaçamentos, através do manejo com regulador de crescimento, foi realizado o presente trabalho, na área experimental da ESALQ/USP, no município de Piracicaba, Estado de São Paulo, no ano agrícola de 2000/01. O delineamento experimental utilizado foi o de blocos ao acaso com parcelas sub-subdivididas e 4 repetições. As parcelas constituíram-se dos espaçamentos de 0,76 e 1,01 m, as subparcelas das três cultivares (IAC 23, DeltaOpal e CD 401) e as sub-subparcelas em presença ou ausência de regulador de crescimento. Em seis plantas marcadas de cada parcela foram analisados os seguintes parâmetros: acompanhamento das fases fenológicas (emergência, aparecimento do primeiro botão floral, antese da primeira flor e abertura do primeiro capulho) e caracteres agronômicos de campo (altura de planta, diâmetro do caule, número de ramos vegetativos e frutíferos, conformação da planta, número de capulhos por planta e altura de inserção do primeiro ramo frutífero). Na área útil de cada parcela foi determinada, a produção de algodão em caroço por hectare e a precocidade de colheita. Em amostras de 20 capulhos tomadas aleatoriamente em cada parcela foram realizadas as análises dos caracteres agronômicos de laboratório (massa de 100 sementes, porcentagem de fibras e massa média de um capulho) e das características tecnológicas da fibra (comprimento, uniformidade de comprimento, maturidade, tenacidade e índice micronaire). Os resultados obtidos permitem concluir que não houve interações significativas para as cultivares x espaçamentos x regulador de crescimento, para a produção de algodão em caroço, precocidade de colheita, caracteres agronômicos de laboratório e características tecnológicas das fibras, com exceção da tenacidade. Desta forma, as cultivares avaliadas, independente da arquitetura da planta, apresentaram capacidade de ajustar-se às variações dos espaçamentos testados e aplicação de regulador de crescimento com relação a estas características. Os caracteres agronômicos de campo como: a altura das plantas e o diâmetro de caule, de modo geral, apresentaram médias reduzidas com a redução do espaçamento e aplicação do regulador de crescimento, para as cultivares de desenvolvimento mais vigoroso (IAC 23 e DeltaOpal), não influenciando, porém, a cultivar de plantas morfologicamente mais compactas (CD 401). / The purpose of this work was to study the phenological, morphological and productive behavior of cotton cultivars presenting different plant shapes under different spacing layouts managed with the use of plant growth regulator. The experiment was established in the experimental field of ESALQ/USP, Piracicaba, Sao Paulo, during the 2000/01 growing season and was arranged in a randomized block experimental design with subdivided plots and 4 replications. The plots represented spacing layouts of 0.76 and 1.01 m, subplots consisted of the IAC 23, DeltaOpal and CD 401 cultivars, while the sub-subplots were characterized by the use or absence of growth regulator. Six plants per plot were labeled and analyzed for the phenological phases (emergence, appearance of the first flower bud, anthesis of the first flower, opening of the first cotton boll) and field agronomical characteristics (plant height, stem diameter, number of vegetative and fruit branches, plant shape, number of cotton boll per plant and insertion height of the first fruit branch). The harvest early and raw cotton yield per hectare were determined considering the effective plot area. The agronomical lab characteristics (mass of 100 seeds, fiber percentage and cotton boll mass), as well as the fiber technological characteristics (length, length uniformity, maturity, strength, and micronaire) were analyzed using a sample of 20 bolls taken at random. Based on the results obtained, it can be concluded that there were no significant interactions among cultivars x spacing x growth regulator, as to cotton yield, harvest early, agronomical lab characteristics and fiber technological characteristics, except for stremgth. The analyzed cultivars, regardless of the plant shape, showed good adaptation to the variation in spacing layouts, as well as to the application of growth regulator, considering the analyzed characteristics. The field agronomical characteristics, such as plant height and stem diameter, generally decreased with the reduction in spacing and with the application of growth regulator, considering the most vigorous cultivars (IAC 23 and DeltaOpal). On the other hand, the cultivar presenting a more compact plant (CD 401) pattern was not influenced by spacing or use of growth regulator.

Dans le contexte de la théorie bayésienne et de théorie de la décision, l'estimation d'une densité prédictive d'une variable aléatoire occupe une place importante. Typiquement, dans un cadre paramétrique, il y a présence d’information additionnelle pouvant être interprétée sous forme d’une contrainte. Cette thèse porte sur des stratégies et des améliorations, tenant compte de l’information additionnelle, pour obtenir des densités prédictives efficaces et parfois plus performantes que d’autres données dans la littérature. Les résultats s’appliquent pour des modèles avec données gaussiennes avec ou sans une variance connue. Nous décrivons des densités prédictives bayésiennes pour les coûts Kullback-Leibler, Hellinger, Kullback-Leibler inversé, ainsi que pour des coûts du type $\alpha-$divergence et établissons des liens avec les familles de lois de probabilité du type \textit{skew--normal}. Nous obtenons des résultats de dominance faisant intervenir plusieurs techniques, dont l’expansion de la variance, les fonctions de coût duaux en estimation ponctuelle, l’estimation sous contraintes et l’estimation de Stein. Enfin, nous obtenons un résultat général pour l’estimation bayésienne d’un rapport de deux densités provenant de familles exponentielles. / Abstract: In the context of Bayesian theory and decision theory, the estimation of a predictive density of a random variable represents an important and challenging problem. Typically, in a parametric framework, usually there exists some additional information that can be interpreted as constraints. This thesis deals with strategies and improvements that take into account the additional information, in order to obtain effective and sometimes better performing predictive densities than others in the literature. The results apply to normal models with a known or unknown variance. We describe Bayesian predictive densities for Kullback--Leibler, Hellinger, reverse Kullback-Leibler losses as well as for α--divergence losses and establish links with skew--normal densities. We obtain dominance results using several techniques, including expansion of variance, dual loss functions in point estimation, restricted parameter space estimation, and Stein estimation. Finally, we obtain a general result for the Bayesian estimator of a ratio of two exponential family densities.

The aim of this graduation thesis is to investigate whether it is possible to design a fluid machine that would utilize a flow resulting from different densities of processing medium, to generate electricity. The thesis focuses on the determination of relations that have influence on the efficiency. The parametres influencing the degree of efficiency have been analyzed. The aim is to reach the highest possible degree of efficiency. Depending on the results, a suitable processing medium can be selected. Within the scope of the thesis, an experiment had been proposed that verified the ground principle of the non-isothermal flow occurence. In conclusion, the outcomes are evaluated with the view of a prospective further developement of the device.

<p> Cobequid Bay is a macrotidal embayment with a large intertidal sand bar complex at the east end, located at the head of the Bay of Fundy. Two sand bars were sampled in order to observe variations in impact pit densities on the surfaces of quartz sand grains, and to identify the pattern of grain size distribution over two intertidal bedforms. </p> <p> V pit densities obtained by viewing the quartz grain surfaces at magnifications of 10,000 and 5,000X with a scanning electron microscope indicate a linear correlation with grain size. The number of small V's appears to be consistent for all grain sizes, with the result that V's larger than approximately 1/5 micrometer produce the linear variation noted. Two trends of V pits formation combine to produce a "plateau", in which two grain size classes have similar V pit densities. This plateau corresponds to the break-point between the traction population and the intermittent suspension population. Trend one consists of the bombardment of large, well exposed grains (transported in the traction layer), by smaller saltating grains, while trend two consists of grain impacts between smaller grains in the concentrated "rheological layer". </p> <p> Two different grain size patterns were discovered on the two sampled bedforms. A fining upward trend toward the crest on both the stoss and slip faces was observed on an ebb dominated megaripple on Selmah Bar. The exact opposite pattern was noted on an ebb dominated sandwave from East Noel Bar. On this bedform, an upward coarsening trend toward the crest was discovered on both the stoss and slip faces. These two patterns correspond to the patterns observed by Dalrymple (1977) in his studies of intertidal bedforms. </p> / Thesis / Bachelor of Science (BSc)