Longitudinal data analysis is common in biomedical research area. Generalized estimating equations (GEE) approach is widely used for longitudinal marginal models. The GEE method is known to provide consistent regression parameter estimates regardless of the choice of working correlation structure, provided the square root of n consistent nuisance parameters are used. However, it is important to use the appropriate working correlation structure in small samples, since it improves the statistical efficiency of β estimate. Several working correlation selection criteria have been proposed (Rotnitzky and Jewell, 1990; Pan, 2001; Hin and Wang, 2009; Shults et. al, 2009). However, these selection criteria have the same limitation in that they perform poorly when over-parameterized structures are considered as candidates. In this dissertation, new working correlation selection criteria are developed based on generalized eigenvalues. A set of generalized eigenvalues is used to measure the disparity between the bias-corrected sandwich variance estimator under the hypothesized working correlation matrix and the model-based variance estimator under a working independence assumption. A summary measure based on the set of the generalized eigenvalues provides an indication of the disparity between the true correlation structure and the misspecified working correlation structure. Motivated by the test statistics in MANOVA, three working correlation selection criteria are proposed: PT (Pillai's trace type criterion),WR (Wilks' ratio type criterion) and RMR (Roy's Maximum Root type criterion). The relationship between these generalized eigenvalues and the CIC measure is revealed. In addition, this dissertation proposes a method to penalize for the over-parameterized working correlation structures. The over-parameterized structure converges to the true correlation structure, using extra parameters. Thus, the true correlation structure and the over-parameterized structure tend to provide similar variance estimate of the estimated β and similar working correlation selection criterion values. However, the over-parameterized structure is more likely to be chosen as the best working correlation structure by "the smaller the better" rule for criterion values. This is because the over-parameterization leads to the negatively biased sandwich variance estimator, hence smaller selection criterion value. In this dissertation, the over-parameterized structure is penalized through cluster detection and an optimization function. In order to find the group ("cluster") of the working correlation structures that are similar to each other, a cluster detection method is developed, based on spacings of the order statistics of the selection criterion measures. Once a cluster is found, the optimization function considering the trade-off between bias and variability provides the choice of the "best" approximating working correlation structure. The performance of our proposed criterion measures relative to other relevant criteria (QIC, RJ and CIC) is examined in a series of simulation studies.

Les écoulements autour de structures en mouvement font l'objet de plusieurs travaux numériques et expérimentaux. L'objectif de ce travail de thèse consiste à montrer la pertinence de la combinaison de la pénalisation volumique avec la méthode de lattice Boltzmann (LBM), dans l'étude du mouvement d'obstacles mobiles dans un écoulement, et de leur interaction avec celui-ci. La LBM,qui est simple et précise à mettre en œuvre, a prouvé ces dernières années son efficacité dans le domaine de la mécanique des fluides. Par ailleurs, la méthode de pénalisation volumique consiste à introduire un terme de pénalisation dans l'équation que l'on souhaite résoudre, afin de prendre en compte l'influence de l'obstacle sur le domaine fluide. Comme cette équation est résolue sur l'ensemble du domaine composé du fluide et du solide, les conditions aux limites à l'interface fluide-solide sont appliquées de façon naturelle. Il semble donc aisé de combiner cette technique avec la méthode de lattice Boltzmann. Nous avons dans un premier temps rappelé les notions de base et les principales caractéristiques de la méthode de lattice Boltzmann. On a présenté quelques exemples d'applications sur des cas tests, que nous avons programmés. Ensuite, une étude bibliographique faisant état des différentes approches qui utilisent la LBM dans l'étude des problèmes d'interaction fluide structure (IFS) a été réalisée. Puis, la combinaison de la pénalisation volumique avec la LBM a été testée avec succès sur l'équation de Burgers monodimensionnelle. La validation s'est portée en premier lieu, sur un écoulement autour d'un solide fixe, puis sur un écoulement autour d'une structure dont le mouvement est imposé, et finalement sur un problème d'IFS de type masse-ressort. La méthode développée a été ensuite testée sur les équations de Navier-Stokes, en considérant un fluide incompressible et une structure rigide. La validation s'est portée tout d'abord sur un écoulement autour d'obstacles immobiles (carré et cylindre), puis autour d'un cylindre mobile en oscillations forcées et libres. Enfin, une dernière application a été portée sur un écoulement entre deux plaques mobiles dans un canal. Nous avons montré que pour tous les cas étudiés, l'approche développée donne de bons résultats, elle reproduit avec précision les résultats de référence. / Flows around moving bodies are the subject of several numerical and experimental studies. The work presented in this document deals with the implementation of a volume penalization technique in a lattice Boltzmann model (LBM), in order to compute flows around moving obstacles. The LBM, which is accurate and easy to implement, has been successfully applied in fluid mechanics during the last decades. It was thus chosen in the present work, for flow computation. Furthermore, the volume penalization technique consists in introducing a volume penalization term into the equation that needs to be solved, in order to take into account the influence of the obstacle on the fluid domain. Since this equation is solved on both fluid and solid domains, the boundary conditions at the fluid-solid interface are naturally applied. Hence this technique seems easy to implement in a lattice Boltzmann framework. In the first chapter, the foundations and the main features of the lattice Boltzmann method are recalled, and several test cases that we simulated are presented. The second chapter deals with a literature review of the techniques developed for the simulation of fluid structure interaction problems in combination with the LBM. In the third chapter, the volume penalization method combined with the LBM was first applied to the one dimensional Burgers equation, considering motionless and moving obstacles (forced motion, and coupling between the fluid force calculated with the penalized Burgers equation and the motion of the obstacle). The combination of the volume penalization approach and the LBM was then employed to solve the incompressible NavierStokes equations, for cases of flows past motionless obstacles (flows over a square obstacle, and past a circular cylinder), and past an oscillating cylinder (where forced and free oscillations of the cylinder were simulated). Finally, this method was also applied to a symmetric Couette flow. For all these simulated cases, a good agreement with numerical results obtained with other techniques, and with results found in literature, was obtained.

Dans ce travail de thèse, une méthode de pénalisation en volume pour la simulation d'écoulements magnétohydrodynamiques (MHD) en domaines confinés est présentée. Les équations incompressibles de la MHD résistives sont résolues par le truchement d'un solveur pseudo-spectral parallèlisé. La pénalisation en volume est une méthode de frontières immergées, caractérisée par une grande flexibilité dans le choix de la géométrie de l'écoulement. Dans le cas présent, elle permet d'utiliser des conditions aux limites non-périodiques dans un schéma pseudo-spectral Fourier. La méthode numérique est validée et sa convergence est quantifiée pour des écoulements hydrodynamiques et MHD, en deux et trois dimensions, en comparant les résultats numériques à ceux de la littérature et à des solutions analytiques. Dans un second temps, la génération spontanée de moment cinétique et magnétique est étudiée pour des écoulements MHD confinés 2D et 3D. L'influence du nombre de Reynolds et du rapport des énergies cinétique/magnétique est explorée, ainsi que les différences induites par les conditions aux limites. Le fait que l'axisymétrie des frontières résulte en un terme de pression non-nul est primordial pour engendrer de grandes valeurs du moment cinétique. L'exclusivité de cette auto-organisation aux écoulements 2D est étudiée en considérant la MHD 3D en présence d'un fort champ magnétique axial. La suite est consacrée à la simulation d'un fluide conducteur dans un cylindre avec un forçage magnétique axial et poloidal. En faisant varier l'amplitude du forçage poloidal, différents états dynamiques sont atteints. Enfin, l'effet du nombre de Prandtl sur le seuil des instabilitées est étudié. / A volume penalization method for the simulation of magnetohydrodynamic (MHD) flows in confined domains is presented. Incompressible resistive MHD equations are solved in 3D by means of a parallelized pseudo-spectral solver. The volume penalization technique is an immersed boundary method, characterized by a high flexibility in the choice of the geometry of the considered flow. In the present case, it allows the use of conditions different from periodic boundaries in a Fourier pseudo-spectral scheme. The numerical method is validated and its convergence is assessed for two- and three-dimensional hydrodynamical and MHD flows by comparing the numerical results with those of the literature or analytical solutions. Then, the spontaneous generation of kinetic and magnetic angular momentum is studied for confined 2D and 3D MHD flows. The influence of the Reynolds number and of the ratio of kinetic/magnetic energies is explored, as well as the differences induced by the boundary conditions. The fact that axisymmetric borders introduce a non-zero pressure term in the evolution equation of the angular momentum is essential to generate large values of the angular momentum. It is investigated whether this self-organization is exclusively observed in 2D flows by considering 3D MHD in the presence of a strong axial magnetic field. The last part is devoted to the simulation of a conducting fluid in a periodic cylinder with imposed axial and poloidal magnetic forcing, implying a resulting magnetic field. By varying the amplitude of the poloidal forcing, different dynamical states can be achieved. The effect of the Prandtl number on the threshold of the instabilities is then studied.

Search engine algorithms are constantly improved in order to provide the highest quality and most relevant results for the user. Google algorithm updates called Panda and Penguin have recently been one of the most significant and many websites that use black hat SEO techniques were hit by them. Nowadays, using this search optimization's techniques raises the risk of Google penalty, which in most of the cases cause decrease in traffic from organic search. The aim of this thesis is to describe techniques of search engine optimization (SEO) and their possible impact on website's ranking. The thesis also aims to verify whether Penguin and Panda updates have the impact on websites for which the methods of black hat SEO have been used and specify the result of this impact. In the theoretical part of this thesis it is described function of search engines Seznam.cz and Google, their algorithm change in history and possible factors on which the ranking score is based. Briefly are also described tools, which can be used to identify and analyze possible causes of penalization. After reading this thesis, the reader will have a basic orientation in search engine optimization methods and their appropriate use. Also, he will be able to recognize possible causes of Google penalty and propose an appropriate strategy for its removal.

With advances in high-throughput techniques, the availability of large-scale omics data has revolutionized the fields of medicine and biology, and has offered a better understanding of the underlying biological mechanisms. However, the high-dimensionality and the unknown association structure between different data types make statistical integration analyses challenging. In this dissertation, we develop three dimensionality reduction methods to detect nonlinear association structure using omics data. First, we propose a method for variable selection in a nonparametric additive quantile regression framework. We enforce a network regularization to incorporate information encoded by known networks. To account for nonlinear associations, we approximate the additive functional effect of each predictor with the expansion of a B-spline basis. We implement the group Lasso penalty to achieve sparsity. We define the network-constrained penalty by regulating the difference between the effect functions of any two linked genes (predictors) in the network. Simulation studies show that our proposed method performs well in identifying truly associated genes with fewer falsely associated genes than alternative approaches. Second, we develop a canonical correlation analysis (CCA)-based method, canonical distance correlation analysis (CDCA), and leverage the distance correlation to capture the overall association between two sets of variables. The CDCA allows untangling linear and nonlinear dependence structures. Third, we develop the sparse CDCA (sCDCA) method to achieve sparsity and improve result interpretability by adding penalties on the loadings from the CDCA. The sCDCA method can be applied to data with large dimensionality and small sample size. We develop iterative majorization-minimization-based coordinate descent algorithms to compute the loadings in the CDCA and sCDCA methods. Simulation studies show that the proposed CDCA and sCDCA approaches have better performance than classical CCA and sparse CCA (sCCA) in nonlinear settings and have similar performance in linear association settings. We apply the proposed methods to the Framingham Heart Study (FHS) to identify body mass index associated genes, the association structure between metabolic disorders and metabolite profiles, and a subset of metabolites and their associated type 2 diabetes (T2D)-related genes. / 2023-11-05T00:00:00Z

Nul ne conteste la nature compensatrice de la responsabilité civile. Pourtant, si cette fonction réparatrice n’est guère discutée, les auteurs se séparent aujourd’hui sur le point de savoir si cette fonction doit être la seule assumée par la responsabilité civile. Non qu’il soit contesté qu’elle assume bien d’autres fonctions ; la grande majorité de la doctrine indique en effet, sans toutefois en souligner les ressorts, que les buts poursuivis par la responsabilité sont multiples et relèvent tant de la compensation des préjudices que de la punition et de la dissuasion des comportements dommageables. A l’heure d’une crise de la responsabilité dénoncée par tous, la question touche donc aux rôles que la responsabilité civile doit assumer. Alors que certains auteurs plaident pour un recentrage dur sa mission première d’indemnisation des victimes, de plus en plus de voix plaident au contraire pour l’élargissement de ses fonctions dans un sens à la fois plus répressif mais aussi préventif. Il s’agit donc désormais d’envisager la responsabilité civile comme un véritable outil de gestion des risques. Totalement axée sur l’indemnisation des victimes depuis plusieurs décennies, la responsabilité civile semble paradoxalement mal armée pour répondre à cette nouvelle attente. De plus, un élargissement des fonctions de la responsabilité civile conduit à s’interroger sur la prééminence d’une fonction sur l’autre ou à leur possible conciliation. En réalité, il existe sans conteste un terreau favorable au développement d’une véritable fonction punitive de la responsabilité civile et l’on constate un double mouvement qui nous semble révélateur du rôle régulateur qu’elle peut assumer. En premier lieu, si l’effet punitif diffus que l’on attachait traditionnellement à la condamnation civile est en net recul, sous l’effet du repli de la faute mais aussi des règles gouvernant la réparation, cet effet subsiste néanmoins, comme le démontre notamment l’analyse économique de la responsabilité. Nous avons donc mis en évidence une résistance latente mais bien réelle d’une conception morale de la responsabilité civile servant son objectif normatif. Surtout, nous avons constaté qu’à cet effet secondaire de la responsabilité tend aujourd’hui à se substituer une véritable fonction punitive et dissuasive de la condamnation civile, construite et cohérente. Nous avons donc proposé une lecture renouvelée du fonctionnement de la responsabilité civile après avoir relevé les germes quelque peu épars de cette conception normative de la responsabilité civile : évolution de la notion de préjudice réparable avec notamment l’assouplissement des critères du préjudice réparable en droit des affaires ou l’émergence de nouveaux préjudices comme en matière environnementale ; modification de l’évaluation et des modalités de la réparation ; émergence de la notion de peine privée. Deux méthodes semblent envisageables afin de soutenir la résurgence de la fonction dissuasive et punitive de la responsabilité civile : la première tient à un alourdissement de la condamnation civile, par le jeu de l’évaluation des préjudices réparables mais aussi par le possible recours à la peine privée ; la seconde correspond à la levée des obstacles à l’effet responsabilisant de la responsabilité civile par l’aménagement des mécanismes d’assurance. Le droit des assurances recèle en effet de mécanismes mal connus ou mal exploités qui peuvent également soutenir l’essor de cette fonction punitive de la responsabilité civile. / No one can contest the nature of the compensation brought about by civil liability. However, even though this reparative function is barely discussed, today the authors are divided on whether or not this should be the unique role assumed by civil liability. It cannot be denied that it takes on other roles and fulfills other purposes; the vast majority of the doctrine indicates that in effect, without underlining the outcomes that the objectives of responsibility are multiple and raise issues concerning the compensation of prejudice as well as punishment and the dissuasion of detrimental behavior. At a time when a crisis of responsibility is denounced by all, the question touches on the roles that civil liability must assume. Whereas certain authors plead for a reevaluation concerning the primary function of indemnity to victims, more and more plead the contrary, for an extension of these functions in a direction that is both more repressive and preventative. It is therefore about considering the use of civil liability as a tool in its own right for risk management purposes.Completely centered upon the compensation for victims for several decades, civil liability has, somewhat paradoxically, been ill equipped to face these new questions. Moreover, a diversification of these functions raises questions on the preeminence of the effects of one function on another and their eventual conciliation.In all actuality, there is undeniably a favorable common ground for the development of a truly punitive function of civil liability and one can note that of a double movement that reveals the regulative role that civil liability can assume.Initially, the punitive effect, which is loosely attached to traditional civil condemnation, is clearly receding under the effect of withdrawing faults but also those rules governing reparation. The effect remains nevertheless, as demonstrated by an economic analysis of responsibility. We have therefore highlighted a latent yet very real moral conception of civil liability serving a normative function.Especially noted for this second function of responsibility is that it tends to substitute a true punitive and dissuasive role of civil judgment that is constructive and coherent.We thus proposed a new interpretation of the role of civil liability after having raised the seedlings of its relatively ill scattered normative design : evolution of the notion of reparable prejudice notably with the easing of criteria relating to reparable prejudice in business law or the emergence of new prejudices, for example within an environmental framework.; the modification of the evaluation of reparation ; the emergence of the notion of privatized punishment.Two methods seem possible in order to support the resurgence of the dissuasive and punitive role of civil liability: the first relates to an increase in the consequences of civil condemnation which plays upon the evaluation of reparable prejudices but also the possible recourse to a private penalty; the second corresponds with removing the stumbling blocks which effect responsibility by putting insurance mechanisms in to place. The law relating to the insurance conceals poorly acknowledged or exploited mechanisms that could support a rise in the punitive function of civil liability.

Pour concevoir les futurs réacteurs à fusion nucléaire, une bonne compréhension des mécanismes régissant l'intéraction plasma-paroi est requise. En particulier, il est nécessaire d'estimer quantitativement les flux de chaleurs impactant les matériaux et la contamination du coeur par les impuretés provenant du mur. Dans ce contexte, le code fluide SolEdge2D a été développé pour simuler le transport dans le plasma de bord. L'interaction plasma-paroi est prise en compte grâce à une méthode de pénalisation innovante et originale. Cette méthode permet en particulier de modéliser la géométrie complexe des éléments face au plasma avec une grande flexibilité. En parallèle, une étude plus théorique sur les propriétés du transport dans les milieux faiblement collisionels a été conduite avec les physiciens du groupe CSDC de l'université de Florence. Une généralisation de la loi de Fourier prenant en compte les corrélation spatio-temporelle à longue distance à été obtenue par l'analyse de modèles stochastiques 1D. Cette loi retrouve en particulier la transition entre un régime diffusif à forte collisionalté et un régime balistique à faible collisionalité. / In the scope of designing future nuclear fusion reactors, a clear understanding of the plasma-wall interaction is mandatory. Indeed, a predictive estimation of heat flux impacting the surface and the subsequent emission of impurities from the wall is necessary to ensure material integrity and energy confinement performances. In that perspective, the fluid code SolEdge2D has been developed to simulate plasma transport in the tokamak edge plasma. The plasma-wall interaction is modeled using an innovative penalization technique. This method enables in particular to take complex plasma facing components geometry into account. In parallel to this numerical effort, a theoretical work has been achieved to find appropriate corrections to fluid closures when collisionality drops. The study of stochastic 1D models has been realized in collaboration with physicists from the CSDC group in Florence. A generalized Fourier law taking long range spatio-temporal correlations has been found to properly account for ballistic transport in the low collisional regime. This formulation is expected to be used to model parallel heat flux or turbulent cross-field transport in tokamak plasmas.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-06-22T20:11:29Z No. of bitstreams: 1 francilenebarbosadossantossilva.pdf: 1221545 bytes, checksum: 856c7ab72d52744ef09d0bed1ecbd238 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-08-07T20:11:10Z (GMT) No. of bitstreams: 1 francilenebarbosadossantossilva.pdf: 1221545 bytes, checksum: 856c7ab72d52744ef09d0bed1ecbd238 (MD5) / Made available in DSpace on 2017-08-07T20:11:10Z (GMT). No. of bitstreams: 1 francilenebarbosadossantossilva.pdf: 1221545 bytes, checksum: 856c7ab72d52744ef09d0bed1ecbd238 (MD5) Previous issue date: 2011-08-31 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Estratégias de penalização são muito utilizadas no trato de problemas com restrições. Problemas inerentes a escolha de valores adequados para os termos de penalização di-ficultam a obtenção de resultados confiáveis e robustos na sua aplicação em problemas da otimização estrutural. Técnicas baseadas em modelos de penalização adaptativa tem apresentado relativo sucesso quando aplicadas em conjunto com algoritmos evolucionis-tas. Apresenta-se aqui uma nova alternativa utilizando uma estratégia de lagrangeano aumentado para o trato das restrições do problema de otimização estrutural. Encontra-se na literatura modelos para penalização adaptativa bem como o uso do lagrangeano aumentado em conjunto com algoritmos genéticos geracionais. O objetivo desse trabalho é adaptar um modelo de penalização para um algoritmo genético não gera-cional, bem como criar um algoritmo baseado em lagrangeano aumentado também para o algoritmo não-geracional. Esses algoritmos foram aplicados em estruturas reticuladas, muito utilizadas na construção civil como coberturas de ginásios, hangares, galpões, etc. O desempenho desses tipos de estruturas e funções matemáticas foi analisado com as técnicas de tratamento de restrição apresentadas nesse trabalho. Isso foi feito durante a busca de soluções ótimas na tentativa de minimizar os custos e satisfazer as restrições adequadas para diversas estruturas e funções matemáticas. / Penalty strategies are widely used in dealing with problems with constraints. Problems inherent in the choice of appropriate values for the terms of penalties dificult to obtain reliable and strong results in its application in problems of structural optimization. Techniques based on models of adaptive penalty has shown some success when applied in conjunction with evolutionary algorithms. Here is presented a new alternative using augmented Lagrangian strategy for dealing with the problem of constrained structural optimizations. It is found in the literature models for adaptive penalties as well as the use of the augmented Lagrangian together with generational genetic algorithms. The aim of this work is to adapt a model of penalization for non-generational genetic algorithm, as well as create an algorithm based on augmented Lagrangian as also for a non-generational algorithm. These algorithms were applied to structures, widely used in construction as coverage of gymnasiums, hangars, etc.. The performance of these types of structures and functions was analyzed using mathematical techniques for handling constraints presented in this work. This was done during the search for optimal solutions in an attempt to minimize costs and satisfy the constraints appropriate for various structures and mathematical functions.

No description available.

Motivée par deux problématiques liées aux plasmas de tokamak, cette thèse propose deux méthodes d'approximation numérique pour deux problèmes mathématiques s'y rattachant. D'une part, pour l'étude du transport turbulent de particules, une méthode numérique basée sur les schémas de subdivision est présentée pour la simulation de trajectoires de particules dans un champ de vitesse fortement variable. D'autre part, dans le cadre de la modélisation de l'interaction plasma-paroi, une méthode de pénalisation est proposée pour la prise en compte de conditions aux limites de type Neumann ou Robin. Analysées sur des problèmes modèles de complexité croissante, ces méthodes sont enfin appliquées dans des situations plus réalistes d'intérêt pratique dans l'étude du plasma de bord. / Motivated by two issues related to tokamak plasmas, this thesis proposes two numerical approximation methods for two mathematical problems associated with them. On the one hand, in order to study the turbulent transport of particles, a numerical method based on subdivision schemes is presented for the simulation of particle trajectories in a strongly varying velocity field. On the other hand, in the context of modeling the plasma-wall interaction, a penalization method is proposed to take into account Neumann or Robin boundary conditions. Analyzed on model problems of increasing complexity, these methods are finally applied in more realistic situations of practical interest in the study of the edge plasma.