Dissertations / Theses on the topic 'Correlation Matrix'

Abstract This thesis project is divided in two parts. The first part examines the possibility that correlation matrix estimates based on an outlier sample would contain information about extreme events. According to my findings, such methods do not perform better than simple shrinkage methods where robust shrinkage targets are used. The method tested is especially outperformed when it comes to the extreme events, where a shrinkage of the correlation matrix towards the identity matrix seems to give the best result. The second part is about valuation of skewness in marginal distributions and the penalizing of heavy tails. I argue that it is reasonable to use a degrees of freedom parameter instead of kurtosis and a certain regression parameter, that I develop, instead of skewness due to robustness issues. When minimizing the one period draw-down is our target, the "value" of skewness seems to have a linear relationship with expected returns. Re-valuing of expected returns, in terms of skewness, in the standard Markowitz framework will tend to lower expected shortfall (ES), increase skewness and lower the realized portfolio variance. Penalizing of heavy tails will most times in the same way lower ES, lower kurtosis and realized portfolio variance. The results indicate that the parameters representing higher order moments in some way characterize the assets and also reflect their future behavior. These properties can be used in a simple optimization framework and seem to have a positive impact even on portfolio level

Cette thèse traite de la description de la corrélation électronique et de la spectroscopie dans le cadre de la Théorie de la Fonctionnelle de la Matrice Densité Réduite (RDMFT). Dans la RDMFT, les propriétés de l'état fondamental d'un système physique sont des fonctionnelles de la matrice densité à un corps. Plusieurs approximations à la corrélation électronique ont été proposées dans la littérature. Beaucoup d'entre elles peuvent être reliés au travail de Müller, qui en a proposé une similaire à l'approximation Hartree-Fock mais qui peut produire des nombres d'occupation fractionnaires. Cela n'est pas toujours suffisant, notamment dans les matériaux fortement corrélés. Par ailleurs, l'expression des observables du système en terme de la matrice densité n'est pas toujours connue. Tel est le cas, par exemple, pour la fonction spectrale, qui est liée aux spectres de photoémission. Dans ce cas, il y a des annulations d'erreur entre l'approximation à la corrélation électronique et l'approximation à l'observable, ce qui affaiblit la théorie. Dans cette thèse, nous recherchons des approximations plus précises en exploitant le lien entre les matrices densité et les fonctions de Green. Dans la première partie de la thèse, nous nous concentrons sur la fonction spectrale. En utilisant le modèle de Hubbard, qui peut être résolu exactement, nous analysons les approximations existantes à cette observable et nous soulignons leurs points faibles. Ensuite, à partir de sa définition en terme de la fonction de Green à un corps nous dérivons une expression pour la fonction spectrale qui dépend des nombres d'occupation naturels et d'une énergie efficace qui prend en compte toutes les excitations du système. Cette énergie efficace dépend de la matrice densité à un corps ainsi que des ordres supérieurs. Des approximations simples à cette énergie efficace donnent des spectres précis dans des systèmes modèles dans des régimes à la fois de faible et de forte corrélation. Pour illustrer notre méthode sur les matériaux réels, nous calculons le spectre de photoemission du NiO massif: notre méthode donne une image qualitativement correcte dans la phase antiferromagnétique et dans la phase paramagnétique, contrairement aux méthodes de champ moyen utilisés actuellement, qui donnent un métal dans le dernier cas. La deuxième partie de la thèse est plus explorative et traite des phénomènes dépendant du temps dans la RDMFT. En général, l'évolution temporelle des matrices densité est donnée par la hiérarchie des équations de Bogoliubov-Born-Green-Kirkwood-Yvon (BBGKY), dans lequel l'équation du mouvement de la matrice densité a n corps est donnée en termes de la matrice densité à n+1 corps. La première équation de la hiérarchie relie la matrice densité à un corps à la matrice densité à deux corps. La tâche difficile est de trouver des approximations à la matrice densité à deux corps. Les approximations existantes sont des extensions adiabatiques des approximations de l'état fondamental. Nous explorons cette question en examinant de nouvelles approximations qui nous tirons de la théorie à plusieurs corps (MBPT) basée sur les fonctions de Green ainsi que de la solution exacte du modèle de Anderson à deux niveaux dans son état fondamental. Nos premiers résultats sur le modèle de Anderson soumis à divers champs externes montrent quelques caractéristiques intéressantes, qui suggèrent d'explorer davantage ces approximations aussi sur des systèmes modèles plus grands
This thesis addresses the description of electron correlation and spectroscopy within the context of Reduced Density-Matrix Functional Theory (RDMFT). Within RDMFT the ground-state properties of a physical system are functionals of the ground-state reduced density matrix. Various approximations to electron correlation have been proposed in literature. Many of them, however, can be traced back to the work of Müller, who has proposed an approximation to the correlation which is similar to the Hartree-Fock approximation but which can produce fractional occupation numbers. This is not always sufficient. Moreover, the expression of the observables of the system in terms of the reduced density matrix is not always known. This is the case, for example, for the spectral function, which is closely related to photoemission spectra. In this case there are error cancellations between the approximation to correlation and the approximation to the observable, which weakens the theory. In this thesis we look for more accurate approximations by exploiting the link between density matrices and Green's functions. In the first part of the thesis we focus on the spectral function. Using the exactly solvable Hubbard model as illustration, we analyze the existent approximations to this observable and we point out their weak points. Then, starting from its definition in terms of the one-body Green's function, we derive an expression for the spectral function that depends on the natural occupation numbers and on an effective energy which accounts for all the charged excitations. This effective energy depends on the one-body as well as higher-order reduced density matrices. Simple approximations to this effective energy give accurate spectra in model systems in the weak as well as strong-correlation regimes. To illustrate our method on real materials we calculate the photoemission spectrum of bulk NiO: our method yields a qualitatively correct picture both in the antiferromagnetic and in the paramagnetic phases, contrary to currently used mean-field methods, which give a metal in the latter case. The second part of the thesis is more explorative and deals with time-dependent phenomena within RDMFT. In general the time evolution of the reduced density matrices is given by the Bogoliubov-Born-Green-Kirkwood-Yvon (BBGKY) hierarchy of equations, in which the equation of motion of the n-body reduced density matrix is given in terms of the (n + 1)-body reduced density matrix. The first equation of the hierarchy relates the one-body to the two-body reduced density matrix. The difficult task is to find approximations to the two-body reduced density matrix. Commonly used approximations are adiabatic extension of ground-state approximations. We explore this issue by looking at new approximations derived from Many-Body Perturbation Theory (MBPT) based on Green's functions as well as from the exact solution of the two-level Anderson impurity model in its ground state. Our first results on the two-level Anderson model subjected to various external fields show some interesting and, at the same time, puzzling features, which suggest to explore further these approximations

Matrix Completion problems have been receiving increased attention due to their varied applicability in different domains. Correlation matrices arise often in studying multiple streams of time series data like technical analysis of stock market data. Often some of the values in the matrix are unknown and some reasonable replacements have to be found at the earliest opportunity to avert an unwanted consequence or keep up the pace in the business. After looking to background research related to solving this problem, we propose a new parallel technique that can solve general correlation matrix completion problems over a set of computers connected to a high speed network. We present some of our results where we could reduce the execution time.

The human brain is an extremely powerful pattern recogniser, as well as being capable of displaying amazing feats of memory. It is clear that human memory is associative; we recall information by associating items together so that one may be used to recall another. This model of memory, where items are associated as pairs rather than stored at a particular location, can be used to implement computer memories which display powerful properties such as robustness to noise, a high storage capacity and the ability to generalise. One example of such a memory is the Binary Correlation Matrix Memory (CMM), which in addition to the previously listed properties is capable of operating extremely quickly in both learning and recall, as well as being well suited for hardware implementation. These memories have been used as elements of larger pattern recognition architectures, solving problems such as object recognition, text recognition and rule chaining, with the memories being used to store rules. Clearly, the performance of the memories is a large factor in the performance of such architectures. This thesis presents a discussion of the issues involved with optimising the performance of CMMs in the context of larger architectures. Two architectures are examined in some detail, which motivates a desire to improve the storage capacity and generalisation capability of the memories. The issues surrounding the optimisation of storage capacity of CMMs are discussed, and a method for improving the capacity is presented. Additionally, while CMMs are able to generalise, this capability is often ignored. A method for producing codes suitable for storage in a CMM is presented, which provides the ability to react to previously unseen inputs. This potentially adds a powerful new capability to existing architectures.

Condition monitoring systems for prognostics and diagnostics can enable large and complex systems to be operated more safely, at a lower cost and have a longer lifetime than is possible without them. AURA Alert is a condition monitoring system that uses a fast approximate k-Nearest Neighbour (kNN) search of a timeseries database containing known system states to identify anomalous system behaviour. This search algorithm, AURA kNN, uses a type of binary associative neural network called a Correlation Matrix Memory (CMM) to facilitate the search of the historical database. AURA kNN is evaluated with respect to the state of the art Locality Sensitive Hashing (LSH) approximate kNN algorithm and shown to be orders of magnitude slower to search large historical databases. As a result, it is determined that the standard AURA kNN scales poorly for large historical databases. A novel method for generating CMM input tokens called Weighted Overlap Code Construction is presented and combined with Baum Coded output tokens to reduce the query time of the CMM. These modifications are shown to improve the ability of AURA kNN to scale with large databases, but this comes at the cost of accuracy. In the best case an AURA kNN search is 3.1 times faster than LSH with an accuracy penalty of 4% on databases with 1000 features and fewer than 100,000 samples. However the modified AURA kNN is still slower than LSH with databases with fewer features or more samples. These results suggest that it may be possible for AURA kNN to be improved so that it is competitive with the state of the art LSH algorithm.

The mechanical environment influences vascular smooth muscle cell (VSMC) functions related to the vascular remodeling. However, the relationships are not appropriately addressed by most mechanical models of arteries assuming homogeneity. Accounting for the effects of heterogeneity is expected to be important to our understanding of VSMC functions. We hypothesized that local stresses computed using a heterogeneous mechanical model of arteries positively correlate to the levels of matrix metalloproteinase (MMP)-2 and -9 in situ. We developed a mathematical model of an arterial wall accounting for nonlinearity, residual strain, anisotropy, and structural heterogeneity. The distributions of elastin and collagen fibers, quantified using their optical properties, showed significant structural heterogeneity. Anisotropy was represented by the direction of collagen fibers, which was measured by the helical angle of VSMC nuclei. The recruiting points of collagen fibers were computed assuming a uniform strain of collagen fibers under physiological loading conditions; an assumption motivated by the morphology. This was supported by observed uniform length and orientation of VSMC nuclei under physiological loading. The distributions of circumferential stresses computed using both heterogeneous and corresponding homogeneous models were correlated to the distributions of expression and activation of MMP-2 and -9 in porcine common carotid arteries, which were incubated in an ex vivo perfusion organ culture system under either normotensive or hypertensive conditions for 48 hours. While strains computed using incompressibility were identical in both models, the heterogeneous model, unlike the homogeneous model, predicted higher circumferential stresses in the outer layer. The tissue levels of MMP-2 and -9 were positively correlated to circumferential stresses computed using the heterogeneous model, which implies that areas of high stress are expected to be sites of localized remodeling and agrees with results from cell culture studies. The results support the role of mechanical stress in vascular remodeling and suggest the importance of structural heterogeneity in studying mechanobiological responses.

In many areas of science one often has a given matrix, representing for example a measured data set and is required to find a matrix that is closest in a suitable norm to the matrix and possesses additionally a structure, inherited from the model used or coming from the application. We call these problems structured matrix nearness problems. We look at three different groups of these problems that come from real applications, analyze the properties of the corresponding matrix structure, and propose algorithms to solve them efficiently. The first part of this thesis concerns the nearness problem of finding the nearest k factor correlation matrix C(X) = diag(I_n -XX T)+XX T to a given symmetric matrix, subject to natural nonlinear constraints on the elements of the n x k matrix X, where distance is measured in the Frobenius norm. Such problems arise, for example, when one is investigating factor models of collateralized debt obligations (CDOs) or multivariate time series. We examine several algorithms for solving the nearness problem that differ in whether or not they can take account of the nonlinear constraints and in their convergence properties. Our numerical experiments show that the performance of the methods depends strongly on the problem, but that, among our tested methods, the spectral projected gradient method is the clear winner. In the second part we look at two two-sided optimization problems where the matrix of unknowns Y ε R {n x p} lies in the Stiefel manifold. These two problems come from an application in atomic chemistry where one is looking for atomic orbitals with prescribed occupation numbers. We analyze these two problems, propose an analytic optimal solution of the first and show that an optimal solution of the second problem can be found by solving a convex quadratic programming problem with box constraints and p unknowns. We prove that the latter problem can be solved by the active-set method in at most 2p iterations. Subsequently, we analyze the set of optimal solutions C}= {Y ε R n x p:Y TY=I_p,Y TNY=D} of the first problem for N symmetric and D diagonal and find that a slight modification of it is a Riemannian manifold. We derive the geometric objects required to make an optimization over this manifold possible. We propose an augmented Lagrangian-based algorithm that uses these geometric tools and allows us to optimize an arbitrary smooth function over C. This algorithm can be used to select a particular solution out of the latter set C by posing a new optimization problem. We compare it numerically with a similar algorithm that ,however, does not apply these geometric tools and find that our algorithm yields better performance. The third part is devoted to low rank nearness problems in the Q-norm, where the matrix of interest is additionally of linear structure, meaning it lies in the set spanned by s predefined matrices U₁,..., U_s ε {0,1} n x p. These problems are often associated with model reduction, for example in speech encoding, filter design, or latent semantic indexing. We investigate three approaches that support any linear structure and examine further the geometric reformulation by Schuermans et al. (2003). We improve their algorithm in terms of reliability by applying the augmented Lagrangian method and show in our numerical tests that the resulting algorithm yields better performance than other existing methods.

Correlational techniques are important tools in multivariate behavioural and social science exploratory research. A wide array of procedures have been proposed for testing (a) whether any of the variables are related, and (b) which variables are related. In the current study, the performance of the procedures currently available for testing these distinct questions is assessed on the primary Neyman-Pearson criterion for an optimal test. According to this criterion, an optimal procedure is the most powerful procedure that controls experimentwise Type I error rate at or below the nominal level. The findings of the first part of this study addressing how to test complete multivariate independence suggest that the statistic traditionally used (QBA) is not the optimal test, and that one of several recently derived statistics (QSE> QSA> QF) should be used. Computational efficiency of the procedures is also considered with the resulting recommendation of the use of QSA- The second part of this study addresses how to test which variables are correlated; the findings suggest the use of a multi-stage order statistics approach with z-tests (CF). The conditions necessary to ensure maximal power when addressing these questions are also considered.
Arts, Faculty of
Psychology, Department of
Graduate

Accurate three-dimensional stress-strain constitutive properties are essential to understanding complex deformation and failure mechanisms for materials with highly anisotropic mechanical properties. The large number of different methods and specimen types currently required to generate three-dimensional allowables for structural design slow down the material characterization. Also, some of the material constitutive properties are never measured due to prohibitive cost of the specimens needed. A method for measurement of three-dimensional constitutive properties using short-beam specimens subject to three-point bend load has been recently developed. This method is based on the Digital Image Correlation (DIC) full-field deformation measurement and closed-form stress approximation. The purpose of this work is to improve the accuracy of the constitutive properties through accurate stress solution. A method is developed based on a combination of full-field strain measurement and nonlinear finite element stress analysis in the material characterization. The nonlinear shear stress-strain relations are the major concern in this work. An iterative procedure is applied to update the nonlinear shear properties using iterative finite element simulations. The accuracy of the numerical procedure is verified by comparing the finite element strain results with full-field measurements. The procedure is further verified using the V-notched beam test results. Excellent agreement has been achieved in the verification. Simplicity of the short-beam specimens and accuracy of the constitutive property approximations make the present method attractive for measurement of three-dimensional stress-strain relations for anisotropic materials at various load rates.

Over the last decades, there is a constant experimentation on new organizational models and structures by many international companies aiming to overcome the difficulties of a highly competitive global market and satisfy the needs of the customers 24/7. As such, the companies have commonly adopted virtual team strategies using also the technological advancements. In order to optimally use their resources, they have also used matrix organizational structures. In such complex and demanding organizational models, the virtual managers face a lot of challenges in order to effectively lead their teams. Therefore, the required managerial competencies and skills related to an effective team performance are of great scientific interest. The objective of this thesis is twofold. Firstly, this thesis aims to examine how managerial competencies of a virtual manager in matrix organizations influence the performance of a team and what the most important skills of these competencies are. Secondly, it aims to identify the managerial skills that a virtual functional manager and a virtual project leader need in order to effectively manage their teams and compare these skillsets. This study adapts a survey strategy with the use of an online questionnaire and gathers responses from professional engineers that virtually work in matrix organizational schemes mainly based in Sweden and Switzerland. The descriptive analysis of the collected data shows that both type of virtual managers (functional and project) have similar managerial skills which are commonly found in managers working in more traditional organizational models. This is not in line with the findings of our study, since in our analysis we considered the virtuality dimension, for which our results show that it strongly affects the virtual team performance. Therefore, the required managerial competence is formed with skills that are more important in the virtual context than in other traditional models. A second finding is that a virtual functional manager requires different skills compared to the virtual project leader, which suggests that the companies should account for the management role in the hiring process of a specific type of virtual manager. A future research on virtual teams in matrix organizations taking into consideration the different levels of virtuality or broader sampling techniques or case studies per industry would be of a great interest.

Mestrado em Matemática Aplicada à Economia e Gestão
Neste trabalho vai-se proceder ao estudo e à aplicação prática sobre sucessões cronológicas reais dos modelos vectoriais ARMA. Estes modelos generalizam os modelos univariados ARMA e os modelos multivariados de função transferência, tendo vantagem sobre estes últimos porque permitem a análise conjunta de sucessões cronológicas que apresentam efeito de feedback. E de esperar que a modelação conjunta de sucessões potencie a capacidade de as descrever, obtendo-se ganhos significativos em termos previsionais. Deste modo, procerder-se-á ao estudo, com base na análise de dois exemplos concretos, do comportamento dos modelos vectoriais ARMA, conffontando-os com os resultados obtidos pelos modelos univariados e pelos modelos de função transferência.
The aim of this work is to present the methodology of the vectorial ARMA models applied to real time series. These models are generalisations of the univariate ARMA models and of the multivariate transfer function models. The advantage of the vectorial ARMA modelling is to allow the joint analysis of the time series which exhibit feedback effects. It is our intention to show that this joint modelization increases the capacity of describing and forecasting. The application was made with the use of two real examples comparing the results ffom the vectorial ARMA, the univariate and the transfer function modelling.
info:eu-repo/semantics/publishedVersion

In this paper we discuss an algorithm for the construction of D-optimal experimental designs for the parameters in a regression model when the errors have a correlation structure. We show that design points can collapse under the presence of some covariance structures and a so called nugget can be employed in a natural way. We also show that the information of equidistant design on covariance parameter is increasing with the number of design points under exponential variogram, however these designs are not D-optimal. Also in higher dimensions the exponential structure without nugget leads to collapsing of the D-optimal design when also parameters of covariance structure are of interest. However, if only trend parameters are of interest, the designs covering uniformly the whole design space are very efficient. For illustration some numerical examples are also included. (author's abstract)
Series: Research Report Series / Department of Statistics and Mathematics

Além das funções mecânicas, sabe-se que o colágeno do tipo I é um dos componentes mais ativos na modulação das atividades biológicas que ocorrem na matriz extracelular (MEC) tais como a adesão celular, a coagulação e a mineralização do osso. Estes eventos ocorrem nos limites do período D, que é provavelmente o domínio funcional que mais prevalece em relação à transferência da informação extracelular para o interior da célula. O objetivo deste trabalho foi estudar as correlações existentes entre as atividades biológicas da MEC e as características estruturais do colágeno do tipo I, principalmente em relação à distribuição topográfica dos aminoácidos (aa) ácidos, básicos e hidrofóbicos. A metodologia empregada foi baseada em informações já existentes sobre a estrutura colágeno do tipo I e suas atividades relacionadas com a MEC. Os resultados mostraram que aa ácidos, básicos e hidrofóbicos formam duas regiões distintas. Em uma delas, aa ácidos e básicos se alojam em áreas delimitadas por aa hidrofóbicos; na outra, aa básicos, ácidos e hidrofóbicos formam aglomerados localizados principalmente nas proximidades dos limites Overlap:Gap. Exceto para os eventos de natureza estrutural são as regiões do colágeno do tipo I mais envolvidas na modulação das atividades da MEC, com a interface Overlap:Gap funcionando como um divisor do controle destas atividades. Enquanto 83,2% das atividades de adesão estão localizadas na zona do Overlap, 93,3% das atividades de natureza estruturais ocorrem a partir da zona limite Overlap:Gap e em direção à zona Gap. Embora nenhuma correlação possa ter sido estabelecida para a resultante de carga, sugere-se que a participação do colágeno do tipo I nas atividades da MEC, possa estar associada a diferenças de interações dos ligantes (carga, massa molecular e hidrofobicidade) com gradientes eletrostático e hidrofóbico existentes ao longo do período D.
Besides mechanical function, it is known today that type I collagen plays a major role in the modulation of many of the activities occurring in the extra-cellular matrix (ECM) such as cell adhesion, coagulation cascade and mineral deposition during bone formation. All these events occur by interaction of cells and ligands within the D-period of collagen microfibrill assembly, which is perhaps the most prevalent module of ECM information in the body. The purpose of this work was the study of correlations between the biological activities of ECM and the structural features of collagen associated with the distribution of acidic, basic and hydrophobic amino acids (aa) within the D period. The methodology used was based on existing information with respect to the structure of type I collagen and its function with respect to ECM activities. The results showed that acidic, basic and hydrophobic aa are found in two major distinct regions: one in which acidic and basic aa are located in areas defined by hydrophobic aa barriers; the other in which acidic, basic and hydrophobic aa forming clusters located preferentially in the boundaries of the Overlap:Gap zones. Except for the structural event, these clusters are the major sites for the interaction involving cell adhesion. The results also suggest that the Overlap:Gap interface function as a borderline for collagen interactions with ECM since 83,2% of the activities related to cell adhesion are located in the Overlap region while 93,3% of the activities of structural nature are enclosed within the limiting zone of Overlap:Gap to the direction of the Gap zone. Although no correlation could be established for net charge along the D period it is suggested that type I collagen participation in the activities of the ECM may also be associated to differences in ligand interactions (charge, molecular mass and hydrophobicity) with the electrostatic and the hydrophobic gradients that exists within the D period.

Os processos espaço-temporais vêm ganhando destaque nos últimos anos, em razão do aumento de estudos compreendendo variáveis que apresentam interação entre as dimensões espacial e temporal. Com o objetivo de modelar esses processos, Pfeifer e Deutsch (1980a) propuseram uma extensão da classe de modelos univariados de Box-Jenkins, denominada por modelo espaço-temporal autoregressivo de média móvel (STARMA). Essa classe de modelos é utilizada para descrever dados de séries temporais espacialmente localizadas. Os processos passíveis de modelagem via classe de modelos STARMA são caracterizados por observações de variáveis aleatórias, em que os locais a serem incorporados no modelo são fixos no espaço. A dependência entre as n séries temporais é modelada por meio da matriz de ponderação, de modo que os modelos da classe STARMA expressem cada observação no tempo t e na localização i como uma média ponderada de combinações lineares das observações anteriores e a inovação defasada no espaço e no tempo conjuntamente. Dada a nova classe de modelos, os objetivos deste estudo foram apresentar a classe de modelos STARMA, implentar computacionalmente, no software R, rotinas que permitam a análise de dados espaço-temporais, com as rotinas implementadas estabelecer e testar modelos de séries temporais aos dados de temperaturas mínimas médias mensais de 8 estações meteorológicas situadas no Paraná e comparar a classe de modelos STARMA com a classe de modelos univariados proposta por Box e Jenkins (1970). Com este estudo verificou-se que na apresentação da classe de modelos STARMA há complexidade no conceito de ordens de vizinhança e na identificação dos modelos espaço-temporais. Em relação a criação de rotinas responsáveis pelas análises de dados espaço-temporais observou-se dificuldades em sua implementação, principalmente no momento de estimação dos parâmetros. Na comparação da classe de modelos STARMA, multivariada, com a classe de modelos SARIMA, univariada, constatou-se que ambos os modelos se ajustaram de maneira satisfatória aos dados, produzindo previsões acuradas.
Spatio-temporal processes have been highlighted lately, due to the increase of studies approaching variables that present interactions between the spatial and temporal dimensions. In order to model these processes, Pfeifer e Deutsch (1980a) have suggested an extension of the Box-Jenkins univariate model class, named spatio-temporal autoregressive moving-average model (STARMA). This model class is used to describe spatially located time series data. The processes prone to be modeled via the STARMA model class are characterized by observations of random variables, in which the locations to be incorporated in the model are spatially fixed. The dependence between the n time series is modeled through the weighing matrix. So STARMA models express each observation at time t and location i as a weighed mean of linear combinations of the previous observations and the jointly lagged innovation in space and time. Given the new class models, the objectives of this study were to present a class of models STARMA, implentar computationally, in textit R software, routines that allow the analysis of spatio-temporal data with the routines implemented to establish and test models time series data of monthly average minimum temperatures of 8 meteorological stations located in Paraná and compare the class of models STARMA with the class of univariate models proposed by Box e Jenkins (1970). With this study it was found that the presentation of the class of models STARMA no complexity in the concept of ordered neighborhood and identification of spatio-temporal models. Regarding the creation of routines responsible for the analysis of spatio-temporal observed difficulties in its implementation, especially at the time of estimation of parameters. In comparison class STARMA models, multivariate, with the class of SARIMA models, univariate, it was found that both models were adjusted satisfactorily to the data, producing accurate forecasts.

This thesis provides novel methodology for statistical analysis of paired high-dimensional genomic data, with the aimto identify gene interactions specific to each group of samples as well as the gene connections that change between the two classes of observations. An example of such groups can be patients under two medical conditions, in which the estimation of gene interaction networks is relevant to biologists as part of discerning gene regulatory mechanisms that control a disease process like, for instance, cancer. We construct these interaction networks fromdata by considering the non-zero structure of correlationmatrices, which measure linear dependence between random variables, and their inversematrices, which are commonly known as precision matrices and determine linear conditional dependence instead. In this regard, we study three statistical problems related to the testing, single estimation and joint estimation of (conditional) dependence structures. Firstly, we develop hypothesis testingmethods to assess the equality of two correlation matrices, and also two correlation sub-matrices, corresponding to two classes of samples, and hence the equality of the underlying gene interaction networks. We consider statistics based on the average of squares, maximum and sum of exceedances of sample correlations, which are suitable for both independent and paired observations. We derive the limiting distributions for the test statistics where possible and, for practical needs, we present a permuted samples based approach to find their corresponding non-parametric distributions. Cases where such hypothesis testing presents enough evidence against the null hypothesis of equality of two correlation matrices give rise to the problem of estimating two correlation (or precision) matrices. However, before that we address the statistical problem of estimating conditional dependence between random variables in a single class of samples when data are high-dimensional, which is the second topic of the thesis. We study the graphical lasso method which employs an L1 penalized likelihood expression to estimate the precision matrix and its underlying non-zero graph structure. The lasso penalization termis given by the L1 normof the precisionmatrix elements scaled by a regularization parameter, which determines the trade-off between sparsity of the graph and fit to the data, and its selection is our main focus of investigation. We propose several procedures to select the regularization parameter in the graphical lasso optimization problem that rely on network characteristics such as clustering or connectivity of the graph. Thirdly, we address the more general problem of estimating two precision matrices that are expected to be similar, when datasets are dependent, focusing on the particular case of paired observations. We propose a new method to estimate these precision matrices simultaneously, a weighted fused graphical lasso estimator. The analogous joint estimation method concerning two regression coefficient matrices, which we call weighted fused regression lasso, is also developed in this thesis under the same paired and high-dimensional setting. The two joint estimators maximize penalized marginal log likelihood functions, which encourage both sparsity and similarity in the estimated matrices, and that are solved using an alternating direction method of multipliers (ADMM) algorithm. Sparsity and similarity of thematrices are determined by two tuning parameters and we propose to choose them by controlling the corresponding average error rates related to the expected number of false positive edges in the estimated conditional dependence networks. These testing and estimation methods are implemented within the R package ldstatsHD, and are applied to a comprehensive range of simulated data sets as well as to high-dimensional real case studies of genomic data. We employ testing approaches with the purpose of discovering pathway lists of genes that present significantly different correlation matrices on healthy and unhealthy (e.g., tumor) samples. Besides, we use hypothesis testing problems on correlation sub-matrices to reduce the number of genes for estimation. The proposed joint estimation methods are then considered to find gene interactions that are common between medical conditions as well as interactions that vary in the presence of unhealthy tissues.

The increasing use of Fiber Reinforced methods for strengthening existing brick masonry walls and columns, especially for the rehabilitation of historical buildings, has generated considerable research interest in understanding the failure mechanism in such systems. This dissertation is aimed to provide a basic understanding of the behavior of solid brick masonry walls unwrapped and wrapped with Fiber Reinforced Cementitious Matrix Composites. This is a new type of composite material, commonly known as FRCM, featuring a cementitious inorganic matrix (binder) instead of the more common epoxy one. The influence of the FRCM-reinforcement on the load-carrying capacity and strain distribution during compression test will be investigated using a full-field optical technique known as Digital Image Correlation. Compression test were carried on 6 clay bricks columns and on 7 clay brick walls in three different configuration, casted using bricks scaled respect the first one with a ratio 1:2, in order to determinate the effects of FRCM reinforcement. The goal of the experimental program is to understand how the behavior of brick masonry will be improved by the FRCM-wrapping. The results indicate that there is an arching action zone represented in the form of a parabola with a varying shape according to the used configuration. The area under the parabolas is considered as ineffectively confined. The effectively confined area is assumed to occur within the region where the arching action had been fully developed.

We consider the finite two-dimensional Ising model on a lattice with periodic boundaryconditions. Kaufman determined the spectrum of the transfer matrix on the finite,periodic lattice, and her derivation was a simplification of Onsager's famous result onsolving the two-dimensional Ising model. We derive and rework Kaufman's resultsby applying representation theory, which give us a more direct approach to computethe spectrum of the transfer matrix. We determine formulas for the spin correlationfunction that depend on the matrix elements of the induced rotation associated withthe spin operator. The representation of the spin matrix elements is obtained byconsidering the spin operator as an intertwining map. We wrap the lattice aroundthe cylinder taking the semi-infinite volume limit. We control the scaling limit of themulti-spin Ising correlations on the cylinder as the temperature approaches the criticaltemperature from below in terms of a Bugrij-Lisovyy conjecture for the spin matrixelements on the finite, periodic lattice. Finally, we compute the matrix representationof the spin operator for temperatures below the critical temperature in the infinite-volume limit in the pure state defined by plus boundary conditions.

The aim of this paper is discussion on particular aspects of the extension of a classic example in the design of experiments under the presence of correlated errors. Such extension allows us to study the effect of the correlation range on the design. We discuss the dependence of the information gained by the D-optimum design on the covariance bandwidth and also we concentrate to some technical aspects that occurs in such settings. (author's abstract)
Series: Research Report Series / Department of Statistics and Mathematics

There are around 10,426 bird species around the world. Recognizing the bird species for an untrained person is almost impossible either by watching or listening them. In order to identify the bird species from their sounds, there is a need for an application that can detect the bird species from its sound. Time-frequency domain analysis techniques are used to implement the application. We implemented two time-frequency domain feature extraction methods. In feature extraction, a signature matrix which consist of extracted features is created for bird sound signals. A database of signature matrix is created with bird chirps extracted features. We implemented two feature classification methods. They are auto-correlation feature classification method and reference difference feature classification method. An unknown bird chirp is compared with the database to detect the species name. The main aim of the research is to implement the time-frequency domain feature extraction method, create a signature matrix database, implement two feature classification methods and compare them. At last, bird species were identified in the research and the auto-correlation classification method detects the bird species better than the reference difference classification method.

The paper considers systems of linear first-order ODEs with a randomly perturbed system matrix and stationary additive noise. For the description of the long-term behavior of such systems it is necessary to study their stationary solutions. We deal with conditions for the existence of stationary solutions as well as with their representations and the computation of their moment functions. Assuming small perturbations of the system matrix we apply perturbation techniques to find series representations of the stationary solutions and give asymptotic expansions for their first- and second-order moment functions. We illustrate the findings with a numerical example of a scalar ODE, for which the moment functions of the stationary solution still can be computed explicitly. This allows the assessment of the goodness of the approximations found from the derived asymptotic expansions.

In 2007, Domínguez-Molina et al. obtained the moment generating function (mgf) of the matrix variate closed skew normal distribution. In this paper, we use their mgf to obtain the first two moments and some additional properties of quadratic forms for the matrix variate skew normal distributions. The quadratic forms are particularly interesting because they are essentially correlation tests that introduce a new type of orthogonality condition.

The rehabilitation of existing masonry elements through jacketing of columns using composite materials is becoming a remarkable technique in several applications that aim to increase the strength of existing masonry building. An experimental campaign was conducted with Steel- and -Basalt Fiber Reinforced Cementitious Matrix (FRCM) systems, in order to test new products that might increase the advantages in terms of good adhesion to masonry substrate, breathability of the system, efficiency in aggressive environments, ease of installation and reversibility, which are essential for the preservation of historical buildings. The mean objective of this experimental study was to investigate the state of the improvement of square masonry columns, built in alternate stretcher and header bond configuration using as material confinement Steel- and-Basalt FRCM system, subjected to axial compression. Moreover, the effectiveness and influence of the confinement in terms of load-bearing capacity and strain distribution with respect to unconfined prisms was carried out. An optical technique, known as Digital Image Correlation (DIC), was employed to understand the interaction between the unit masonry components.

O trabalho tem por objetivo avaliar diferentes métodos para identificar zeros na matriz de transferência de sistemas MIMO e propor um método novo baseado em análise de correlação. Estes métodos são utilizados durante a etapa de pré-identificação, a fim de se obter informações relevantes que possam ser utilizadas para se reduzir o tempo dos experimentos, diminuir a variabilidade dos parâmetros dos modelos e melhorar a eficácia dos modelos remanescentes. Estes métodos são aplicados a sistemas MIMO lineares, com dados coletados em malha aberta e em malha fechada. É avaliado o ganho obtido em relação à capacidade de predição dos modelos, a redução do tempo de identificação e o ganho de desempenho do controlador MPC que utiliza estes modelos. O trabalho conclui que a informação de zeros resulta em melhorias no tempo de identificação e no desempenho do controlador MPC.
This work aims to evaluate different methods to identify zeros in the transfer matrix of MIMO systems and to propose a new method based on correlation analysis. These methods are used during the pre-identification stage in order to identify relevant information that can be used to reduce the duration of the experiment, decrease model parameter variability and improve the accuracy of the remaining models. These methods are applied to MIMO linear systems, with data collected in open and closed-loop. The gains obtained in relation to the predictive ability of the models, the reduction of identification time and the performance gain of the MPC using these models are evaluated. This work concludes that the zero information results in improvements in identification time and in performance gain of the MPC controller.

Considering the crucial involvement of matrix metalloproteinases’ (MMPs) misregulated activity in the pathogenesis of several degenerative diseases, this class of enzymes has been considered a highly active set of targets for the design of new therapeutic agents. However, the scant success of synthetic MMP inhibitors, largely due to the disappointing results obtained in both clinical and preclinical studies, makes medicinal plants a valuable source of new active compounds able to modulate MMPs activity. In this work, a consistent number of plants, selected on the base of an ethnobotanical research, were tested as inhibitors of collagenase, the founding member of the MMPs family. 1H-NMR-based metabolomic analysis combined with multivariate data treatment (PLS and OPLS) was used to correlate the biological activity to the phytochemical profiles, suggesting tannins as an important class of collagenase inhibitors. Thus, a tannin-removal procedure was developed, which allowed to prove this hypothesis and to identify another class of active metabolites, the glucuronide-conjugated flavonoids (especially quercetin-3-O-β-glucoronide), whose the plant Alchemilla vulgaris was found to be a good source. In another stage of the project, different varieties of tea were investigated as collagenase inhibitors, finding black tea samples particularly potent. Then, an OPLS model was developed with the aim of correlating the biological activity to the UV-Vis spectra of teas, showing that a high activity was related to absorption values in the range 350-440 nm. A subsequent fractionation of the most active tea sample was carried out, and this approach allowed to corroborate the results obtained by the metabolomic analysis. Considering that the absorbance measurement of an extract represents a cheap and simple procedure, the proposed method can be suitable, for instance, to select the best tea variety to be developed as an anti-wrinkles cosmetic or food supplement.

In Kenntnis der beschriebenen verfahrenstechnischen Einflussfaktoren auf die Struktur und die Eigenschaften der faserverstärkten MMC geht es darum, ein prepregtechnologisches Verfahren für das Herstellen lang- bzw. endlosfaserverstärkter MMC mit Aluminiummatrix zu erarbeiten und die Machbarkeit im Hinblick auf die Reproduzierbarkeit und das Erreichen der prinzipiell möglichen mechanischen Eigenschaftskennwerte nachzuweisen. Es geht darum, den Einfluss der Prozessparameter auf die Struktur der Prepregs und der verdichteten MMC`s zu erfassen und zu bewerten. In Kenntnis dieser Zusammenhänge sind die qualitätsrelevanten Kennwerte der Werkstoffstruktur denen der mechanischen Eigenschaften gegenüberzustellen. In Weiterführung der o.g. Darstellungen zwischen Struktur- und Eigenschaften der Verbundwerkstoffe geht es darum, die experimentell bestimmten Kennwerte der Festigkeit und des E-Moduls auf Übereinstimmung mit den entsprechenden Korrelationsmodellen (Shear Lag Modell, Grenzwert- und Modellkonzept, EIAS-Methode) zu überprüfen. Ferner geht es darum, die Zusammenhänge zwischen den im Modell definierten idealen Gefügebedingungen mit den realen im Hinblick auf deren Einfluss auf die Eigenschaftskennwerte kritisch zu diskutieren. Aufbauend auf den definierten Zielstellungen kann nach der erfolgten Bewertung aller Untersuchungsergebnisse gezeigt werden, dass die entworfene Technologie zur Herstellung langfaserverstärkter Metallmatrix-Verbundwerkstoffe dazu geeignet ist, Fasern mit geringem Durchmesser und hoher Flexibilität (am Beispiel der hochfesten Kohlenstofffasern vom Typ HTA 5131) mit dem ausgewählten Matrixmetall (am Beispiel der AlSi5-Legierung) zunächst zu Prepregs und weiterhin in verdichtete Verbundstrukturen zusammenzuführen. Das vergleichende Gegenüberstellen von Simulation und Experiment dienen einerseits dem Verifizieren der Prozessmodelle. Die Prozessmodelle stellen andererseits die wertvolle versuchstechnische Grundlage zur Definition und auch der Einengung des Arbeitsfensters für die Prozessparameter und der Anzahl erforderlicher Verifikationsexperimente im Hinblick auf die Probebeschichtungen und deren Auswertung dar. Derartige Modelle leisten eine wichtige Hilfestellung zum weiteren Erhöhen der Prozessstabilität und damit auch dem Nachweis der Reproduzierbarkeit
The aim of this work is a described prepreg-technological method to create aluminium based MMC reinforced with continuous fibres and the verification of reproducibility to achieve relevant mechanical properties. This aim is based on the knowledge of technological influences on structure and properties of fibre reinforced MMC. And so activities are focussed on the evaluation of the influence of process functions on structure and mechanical properties of prepregs and compacted MMC. The comparison between structure and properties is necessary to describe the correlation function of composite material. Furthermore the application of different correlation models (“Shear Lag”, “Grenzwert- Modellkonzept“, “EIAS”) to describe the influence of composite structures on strength and Young`s modulus is necessary to compare theoretical results with those of relevant experiments. The object is a critical quantification of the influence of real structure parameters compared with those of a model defined structure. In view of the described aim it is shown that the described technology to manufacture long fibre reinforced MMC is applicable for preparation of carbon fibres (HTA 5131) with low diameter and high flexibility and metal matrix (AlSi5) to prepregs and compact composites. The comparison of simulated and experimental results is the base for verification of different process models. So it is possible to describe and optimize the process function and moreover to minimize the number of technological experiments. After optimization specified models are a good base to achieve a high level of stability and reproducibility for all steps in prepreg technology

Orientador: Rogerio Custodio
Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica
Made available in DSpace on 2018-08-12T23:54:09Z (GMT). No. of bitstreams: 1 Angelotti_WagnerFernandoDelfino_D.pdf: 827045 bytes, checksum: 044a9aecfaec7ac3e14153480c625d9f (MD5) Previous issue date: 2009
Resumo: Esta tese explorou diferentes objetivos envolvendo o método Monte Carlo Quântico, dos quais se destacam: avaliação do método convencional em propriedades eletrônicas; formalização das relações existentes entre a teoria de matriz densidade e os métodos Monte Carlo Quântico Variacional e de Difusão; estudo da correlação eletrônica com diferentes funções correlacionadas e também através de método misto envolvendo a teoria de perturbação e o Monte Carlo Quântico Variacional; aplicações para átomos do primeiro e segundo período da tabela periódica e moléculas diatômicas. Experimentos computacionais com o método Monte Carlo Quântico e separação de spins foram realizados produzindo excelentes resultados para cálculos de potenciais de ionização sucessivos para átomos, ionização atômica e molecular e construção de curvas de potencial para moléculas simples. Foram ainda obtidas duas formulações analíticas que descrevem exatamente o vínculo formal entre a matriz densidade e o Monte Carlo Quântico. Esta associação proporcionou ótimos resultados para os métodos Variacional e de Difusão, apresentando semelhanças e significativas diferenças quando comparado ao tratamento convencional com respeito à estrutura nodal para cada estado eletrônico estudado. Além disso, a matriz densidade aliada às funções correlacionadas é capaz de recuperar parte da correlação eletrônica e torna possível a correção de funções de onda dentro da associação do Monte Carlo Quântico e teoria de perturbação.
Abstract: This thesis explored different goals involving the quantum Monte Carlo method, of which stand out: assessment of the conventional method in electronic properties; formalization of relations between the density matrix theory and the variational and diffusion quantum Monte Carlo methods; study of the electronic correlation with different correlated functions and also through mixed method involving the perturbation theory and variational quantum Monte Carlo; applications to atoms of the first and second period of the periodic table and diatomic molecules. Computational experiments with the quantum Monte Carlo method and separation of spins were achieved producing excellent results for calculations of successive ionization potentials for atoms, single ionization of atoms and simple molecules and calculation of potential curves for simple molecules. Two analytical formulations were obtained that describes exactly the formal link between the density matrix and quantum Monte Carlo. This association has provided excellent results for variational and diffusion methods, presenting similarities and significant differences when compared to conventional treatment with respect to the nodal structure for each electronic state studied. Furthermore, the density matrix together with correlated wave functions is able to recover part of the electronic correlation and makes possible the correction of the wave functions within the association of quantum Monte Carlo and perturbation theory.
Doutorado
Físico-Química
Doutor em Ciências

Molecular dynamics simulations were performed on the influenza A non-structural protein 1 (NS1) RNA binding domain (RBD), a homodimer. Fourteen simulations were performed at 298K, nine ionized with 0.1M KCl and five with no ions. Several analysis techniques were employed to study RBD residue flexibility. The focus of the study was the RNA binding cavity formed by side chains of helix 2 (chain A) and helix 2’ (chain B) and cavity intermonomeric salt bridges. Opening of the salt bridges D29–R46’ and D29’–R46 was observed in several of the trajectories. The RNA binding cavity has large flexibility, where the dimension and shape change during the dynamics. One pair of residues surrounding the cavity and necessary for RNA binding, residues R38 and R38’, have motions during the simulations which cover the top of the cavity. There is correlation between the salt bridge breaking, flexibility of R38 and R38’, and the cavity size and shape changes. Possible RBD small molecule drug targets are these two salt bridges and the pair R38 and R38’. Disrupting the events that occur around these areas could possibly inactivate RNA binding function of the domain. These results could have implications in searching for potential molecules that effectively treat influenza A.

Les composites à matrice organiques (CMO) et fibre de carbone sont de plus en plus employés dans la réalisation de structures « tièdes » (aubes de fan, nacelles …) ; ces pièces peuvent être soumises, en service, à la fatigue mécanique, au cyclage thermique et à la fatigue thermo-mécanique. Bien qu’il existe une littérature consistante sur le comportement en fatigue des composites tissés, l'interaction entre fatigue et la dégradation liée à l'environnement à haute température n’a pas été encore bien exploitée. Le couplage entre les effets de la thermo-oxydation, le comportement mécanique (viscoélastique, viscoplastique) de la matrice organique à températures élevées et la dégradation par fatigue peut être néfaste pour le composite.Le but de ce travail est de caractériser et de modéliser - pour les composites tissés C/matrice organique - le comportement thermomécanique, l'apparition et le développement de l’endommagement liés aux mécanismes mécaniques cycliques (fatigue) sous environnement contrôlé (température et gaz).Une étude préliminaire sur un composite stratifiée [02/902]s a été menée pour pouvoir analyser les effets de l’environnement sur une architecture simple. La corrélation d’image numérique (CIN) et des scans μ-tomographiques (μCT) ont été employés pour le suivi et la caractérisation de l’endommagement de fatigue de composites tissés 2D à architecture complexe pour applications aéronautiques. Les effets de l’environnement sur la dégradation par fatigue ont été également explorés.L'objectif à long terme de cette étude est de fournir des outils expérimentaux et numériques pour renforcer la compréhension et la modélisation du couplage mécanique/endommagement/environnement pour la prédiction de la durée de vie et pour la proposition de protocoles d’essais accélérés réalistes de pièces « tièdes » en CMO
In the next future, the employment of organic matrix/carbon fibre composites (OMC) is foreseen for the realization of “hot” structures: these parts may be subjected, in service, to mechanical fatigue (e.g. fan blades turbo-engines), thermal cycling and thermo-mechanical fatigue (e.g. aircraft structural parts). Though there is a consistent literature concerning the fatigue behaviour of woven composites, the interaction between fatigue and environmental degradation at high temperature has been poorly explored. Coupling between thermo-oxidation effects, mechanical (viscoelastic, viscoplastic) behaviour of the polymer matrix at high temperatures and degradation due to fatigue may be highly detrimental for the material. This work aims at characterizing and modelling - for carbon fibre/organic matrix (polyimide) textile composites – the thermomechanical behaviour, the onset and the development of damage related to cyclic mechanical mechanisms (fatigue) under controlled (temperature and gas) environment.A preliminary study on a cross-ply laminate [02/902]s has been carried out in order to analyse the environmental effect on a model sample. Digital Image Correlation (DIC) and μ-Computed Tomography (μCT) have been used to monitor and characterize the fatigue damage of 2D woven composites for aeronautical applications. The environmental effect on fatigue degradation have been also explored.The long-term aim of the study is to provide experimental and numerical tools to strengthen the understanding and the modelling of mechanics/damage/environment coupling for durability prediction

Tesis por compendio
[ES] El análisis del control genético del hospedador sobre su microbiota ha sido señalado recientemente como un tema prometedor en diferentes campos de estudio. La relación entre el holobionte hospedador-microbioma y los fenotipos en el ganado lechero podría conducir a nuevos conocimientos en los programas de selección genética. Dentro de esta tesis doctoral, se realizó la estimación y análisis a través de diferentes enfoques estadísticos con el objetivo de desentrañar el control genético del hospedador sobre la microbiota en ganado lechero. Además, se analizó el rasgo de concentración de metano como un fenotipo potencial para ser incluido en el programa de mejora de ganado lechero español. Mayor abundancia relativa de la mayoría de los eucariotas (principalmente protozoos ciliados y hongos) y algunas arqueas (Methanobrevibacter spp. Methanothermus spp. y Methanosphaera spp.) fueron factores de riesgo para ser clasificadas en la categoría alta. Se propuso un conjunto de modelos de ecuaciones estructurales (SEM) de tipo recursivo dentro de un marco de Cadenas de Markov Monte Carlo (MCMC) para analizar conjuntamente la relación hospedador-metagenoma-fenotipo. Se estableció un modelo bivariado no-recursivo como punto de referencia. La heredabilidad de CH4 se estimó en 0,12 ± 0,01 en ambos modelos, recursivo y no recursivo. Asimismo, las estimaciones de heredabilidad para la abundancia relativa de los taxones se superpusieron entre los modelos y variaron entre 0.08 y 0.48. Las correlaciones genéticas entre la composición microbiana y el CH4 variaron de -0,76 a 0,65 en el modelo bivariado no recursivo y de -0,68 a 0,69 en el modelo recursivo. Doce matrices de relación de microbiota (K) fueron construidas a partir de diferentes métricas de distancia del microbioma, con el objetivo de comparar su desempeño dentro de un marco de estimación de componentes de varianza para CH4 y toda la microbiota. Análisis de simulación (n = 1000) y datos reales fueron desarrollados considerando cuatro modelos posibles: un modelo genómico aditivo (GBLUP), un modelo de microbioma (MBLUP), un modelo de efectos genéticos y microbioma (HBLUP) y un modelo de efectos de interacción genético, microbioma y genético × microbioma (HiBLUP). Un nuevo término "Holobiabilidad" fue definido para referirse a la proporción de la varianza atribuible a los efectos del holobionte hospedador-microbioma. Las estimaciones a partir de datos reales usando HiBLUP variaron dependiendo de la K utilizada y estuvieron entre 0.15-0.17, 0.15-0.21 y 0.42-0.59 para heredabilidad, microbiabilidad y holobiabilidad, respectivamente. El conjunto de datos de microbioma fue agregado a través de análisis de componentes principales (PCA), en pocos componentes principales (PCs) que fueron utilizados como aproximaciones del metagenoma central. Parte de la variabilidad condensada en estos PC está controlada por el genoma de la vaca, con estimaciones de heredabilidad para el primer PC (PC1) de ~ 0,30 en todos los niveles taxonómicos, con una gran probabilidad (> 83%) de que la distribución posterior sea > 0,20 y con un intervalo de mayor densidad posterior al 95% (95% HPD) no conteniendo cero. La mayoría de las estimaciones de correlación genética entre PC1 y metano fueron grandes (>0,70) en todos los niveles taxonómicos, con la mayor parte de la distribución posterior (> 82%) siendo > 0,50 y con su 95% HPD no conteniendo cero. Estos resultados sugieren que todo el metagenoma del rumen regula recursivamente las emisiones de metano en las vacas lecheras, y que tanto el CH4 como las composiciones de la microbiota están parcialmente controladas por el genotipo del hospedador. Las variables agregadas (PC) propuestas podrían ser usadas en programas de mejora de animales para reducir las emisiones de metano en las generaciones futuras.
[CA] L'anàlisi del control genètic de l'hoste sobre la seva microbiota s'ha assenyalat recentment com un tema prometedor en diferents camps d'estudi. La relació entre el holobiont hoste-microbioma i els fenotips en bovins de llet podria conduir a nous coneixements en els programes de cria. Dins d'aquest doctorat es van realitzar tesis, estimacions i anàlisis mitjançant diferents enfocaments estadístics amb l'objectiu de desentranyar el control genètic de l'hoste sobre la microbiota en bestiar lleter. A més, es va analitzar el tret de concentració de metà com a fenotip potencial a incloure en el programa espanyol de cria de bestiar lleter. La major abundància relativa de la majoria dels eucariotes (principalment protozous i fongs ciliats) i algunes arquees (Methanobrevibacter spp. Methanothermus spp i Methanosphera spp.) Van ser factors de risc per classificar-se en les categories altes. Es va proposar un conjunt de models d'equacions estructurals (SEM) de tipus recursiu dins d'un marc de cadena Markov Monte Carlo (MCMC) per analitzar conjuntament la relació hoste-metagenoma-fenotip. Es van establir models no recursius com a referència. L'heretabilitat del CH4 es va estimar en 0,12 ± 0,01 en ambdós models, recursius i no recursius. De la mateixa manera, les estimacions d'heretabilitat de l'abundància relativa dels tàxons es van superposar entre models i van oscil·lar entre 0,08 i 0,48. Les correlacions genètiques entre la composició microbiana i el CH4 van oscil·lar entre -0,76 i 0,65 en els models bivariables no recursius i de -0,68 a 0,69 en els models recursius. Dotze matrius de relació de microbiota (K) de diferents mètriques de distància de microbiomes, amb l'objectiu de comparar el seu rendiment dins d'un marc d'estimació de components de variància per CH4 i anàlisi de microbiomes sencers en simulació (n = 1000, 25 rèpliques) i es van realitzar dades reals , considerant quatre possibles models: un model genòmic additiu (GBLUP), un model de microbioma (MBLUP), un model d'efectes genètics i microbiomes (HBLUP) i un model d'efectes d'interacció genètics, microbiomes i genètics × microbiomes (HiBLUP). Es va definir un nou terme "Holobiabilitat" per referir-se a la proporció de la variància fenotípica atribuïble als efectes holobiont del microbioma host. Les estimacions de dades reals mitjançant HiBLUP van variar en funció de la K utilitzada i van oscil·lar entre 0,15-0,17, 0,15-0,21 i 0,42-0,59 per heretabilitat, microbiabilitat i holobiabilitat, respectivament. El conjunt de dades de microbiomes es va agregar mitjançant l'anàlisi de components principals (PCA) en pocs components principals (PC) que es van utilitzar com a proxies del metagenoma principal. Part de la variabilitat condensada en aquestes PC està controlada pel genoma de la vaca, amb estimacions d'heretabilitat per a la primera PC (PC1) de ~ 0,30 a tots els nivells taxonòmics, amb una gran probabilitat (> 83%) de la distribució posterior> 0,20 i amb un 95% més alt interval de densitat posterior (95% HPD) que no conté zero. La majoria de les estimacions de correlació genètica entre PC1 i metà eren grans (>0,70) en tots els nivells taxonòmics, amb una gran part de la distribució posterior (> 82%)> 0,50 i amb un 95% de HPD que no contenia zero. Aquests resultats suggereixen que tot el metagenoma del rumen regula recursivament les emissions de metà en vaques lleteres i que tant el CH4 com les composicions de microbiota estan parcialment controlades pel genotip de l'hoste. Les variables agregades proposades (PC) es podrien utilitzar en programes de cria d'animals per reduir les emissions de metà en les generacions futures.
[EN] The analysis of the host genetic control over its microbiota has recently been pointed out as a promising theme in different fields of study. The relationship between the host-microbiome holobiont and phenotypes in dairy cattle could lead to new insights in breeding programs. Within this Ph.D. thesis, estimation and analysis through different statistical approaches were performed aiming to unravel the host genetic control over the microbiota in dairy cattle. Besides, methane concentration trait was analyzed as a potential phenotype to be included in the Spanish dairy cattle breeding program. Higher relative abundance of most eukaryotes (mainly ciliate protozoa and fungi) and some archaea (Methanobrevibacter spp. Methanothermus spp and Methanosphera spp.) were risk factors for being classified in the high categories. a set of structural equation models (SEMs) of a recursive type within a Markov chain Monte Carlo (MCMC) framework was proposed to jointly analyze the host-metagenome-phenotype relationship. Non-recursive models were set as benchmark. Heritability of CH4 was estimated at 0.12 ± 0.01 in both, the recursive and non-recursive, models. Likewise, heritability estimates for the relative abundance of the taxa overlapped between models and ranged between 0.08 and 0.48. Genetic correlations between the microbial composition and CH4 ranged from -0.76 to 0.65 in the non-recursive bivariate models and from -0.68 to 0.69 in the recursive models. Regardless of the statistical model used, positive genetic correlations with methane were estimated consistently for the 7 genera pertaining to the Ciliophora phylum, as well as for those genera belonging to the Euryarchaeota (Methanobrevibacter sp.), Chytridiomycota (Neocallimastix sp.) and Fibrobacteres (Fibrobacter sp.) phyla. Twelve microbiota relationship matrices (K) from different microbiome distance metrics were built, aiming to compare its performance within a variance component estimation framework for CH4 and whole microbiome analysis on simulation (n = 1000, 25 replicates) and real data were performed, considering four possible models: an additive genomic model (GBLUP), a microbiome model (MBLUP), a genetic and microbiome effects model (HBLUP) and a genetic, microbiome and genetic × microbiome interaction effects model (HiBLUP). A new term "Holobiability" was defined to refer to the proportion of the phenotypic variance attributable to the host-microbiome holobiont effects. Estimates from real data using HiBLUP varied depending on the K used and ranged between 0.15-0.17, 0.15-0.21 and 0.42-0.59 for heritability, microbiability and holobiability, respectively. The microbiome dataset was aggregated through Principal Component Analysis (PCA) into few principal components (PCs) that were used as proxies of the core metagenome. Part of the variability condensed in these PCs is controlled by the cow genome, with heritability estimates for the first PC (PC1) of ~0.30 at all taxonomic levels, with a large probability (>83%) of the posterior distribution being > 0.20 and with the 95% highest posterior density interval (95%HPD) not containing zero. Most genetic correlation estimates between PC1 and methane were large (>0.70) at all taxonomic levels, with most of the posterior distribution (>82%) being >0.50 and with its 95%HPD not containing zero. These results suggest that rumen's whole metagenome recursively regulate methane emissions in dairy cows, and that both CH4 and the microbiota compositions are partially controlled by the host genotype. The purposed aggregated variables (PCs) could be used in animal breeding programs to reduce methane emissions in future generations.
This research was financed by RTA2015-00022-C03-02 (METALGEN) project from the national plan of research, development and innovation 2013-2020 and the Department of Economic Development and Competitiveness (Madrid, Spain). We thank the regional Holstein Associations and farmers collaborating in the project. Computational support from the High-Performance Computing Centre in Galicia (Spain) is acknowledged. Alejandro Saborío-Montero acknowledges the scholarship from Universidad de Costa Rica for his doctorate studies which partially conducted to the progress of this study.
Saborío Montero, A. (2021). Study of the Host Genetic Control over the Ruminal Microbiota and their Relationships with Methane Emissions in Dairy Cattle [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/172633
TESIS
Compendio

Data are valuable assets to any organizations or individuals. Data are sources of useful information which is a big part of decision making. All sectors have potential to benefit from having information. Commerce, health, and research are some of the fields that have benefited from data. On the other hand, the availability of the data makes it easy for anyone to exploit the data, which in many cases are private confidential data. It is necessary to preserve the confidentiality of the data. We study two categories of privacy: Data Value Hiding and Data Pattern Hiding. Privacy is a huge concern but equally important is the concern of data utility. Data should avoid privacy breach yet be usable. Although these two objectives are contradictory and achieving both at the same time is challenging, having knowledge of the purpose and the manner in which it will be utilized helps. In this research, we focus on some particular situations for clustering and classification problems and strive to balance the utility and privacy of the data. In the first part of this dissertation, we propose Nonnegative Matrix Factorization (NMF) based techniques that accommodate constraints defined explicitly into the update rules. These constraints determine how the factorization takes place leading to the favorable results. These methods are designed to make alterations on the matrices such that user-specified cluster properties are introduced. These methods can be used to preserve data value as well as data pattern. As NMF and K-means are proven to be equivalent, NMF is an ideal choice for pattern hiding for clustering problems. In addition to the NMF based methods, we propose methods that take into account the data structures and the attribute properties for the classification problems. We separate the work into two different parts: linear classifiers and nonlinear classifiers. We propose two different solutions based on the classifiers. We study the effect of distortion on the utility of data. We propose three distortion measurement metrics which demonstrate better characteristics than the traditional metrics. The effectiveness of the measures is examined on different benchmark datasets. The result shows that the methods have the desirable properties such as invariance to translation, rotation, and scaling.

Les transitions gamma qui peuplent ou desexcitent la bande yrast superdeformee du noyau 132 cerium relient des etats qui se trouvent a une energie d'excitation elevee, ou la densite d'etats importante qui lui est associee rend difficile leur observation discrete. L'etude porte donc sur la composante continuum des spectres gamma. Les donnees analysees proviennent de la reaction #1#0#0mo(#3#6s,4n)#1#3#2ce, a une energie de bombardement de 155 mev. Elles ont ete recueillies par le multidetecteur eurogam dans sa phase 1 et 2, et l'amelioration de la qualite des donnees entre ces deux phases est appreciee. Les evenements enregistres ont ete tries et seulement ceux qui contiennent au minimum 2 transitions de la bande rotationnelle superdeformee yrast ont ete stockes sur disque. Des spectres fortement correles a la bande ont ainsi ete construits. Une methode de soustraction de fond des matrices de correlation gamma-gamma a ete developpee, et son utilisation a permis de mettre en evidence le peuplement de la bande yrast superdeformee par un continuum rotationnel, correspondant a une forme tres allongee du noyau. L'evolution des correlations rotationnelles de ce continuum a pu etre caracterisee. A l'aide des evenements d'ordre de coincidence eleve recueillis par le multidetecteur, une strategie d'analyse a permis d'isoler une composante du spectre correspondant a la desexcitation de la bande yrast superdeformee. Le centroide ainsi que la multiplicite de cette distribution permettent d'evaluer l'energie d'excitation de la bande yrast superdeformee par rapport aux etats normalement deformes du noyau

Cette thèse a pour objectif de développer une méthode de mesure de champs par corrélation d’images numériques (CIN) à haute température couplée à des mesures thermiques sur une éprouvette technologique en CMC sollicitée dans des conditions thermiques représentatives d’un environnement moteur et de mettre en place une méthodologie d’identification des propriétés thermiques et thermomécaniques du matériau, en quantifiant à chaque étape de la chaîne les incertitudes associées aux quantités d’intérêt et en les réduisant. Il a pour cela été nécessaire de traiter les défis inhérents à la CIN à chaud, que ce soit au niveau de l’acquisition des images (saturation, perte du contraste) ou de la mesure (artefacts dus à l’effet mirage, aussi appelé "brume de chaleur").Ces travaux ont ainsi donné lieu au développement d’un protocole d’étalonnage d’un banc multi-instrumenté par l’utilisation soit d’une mire in-situ, soit par auto-étalonnage en utilisant l’éprouvette elle-même et son environnement. Les mesures de déplacements 3D surfaciques (approches de stéréocorrélation globales) et les mesures thermiques ont permis de mettre en évidence ce phénomène de brume de chaleur. Des stratégies de régularisation spatiotemporelles des déplacements mesurés ont été proposées et ont permis d’obtenir des résultats satisfaisants (réduction significative des incertitudes de mesure). De même, des approches par réduction de modèles (POD) ont permis de traiter les données thermiques et de quantifier les incertitudes associées aux phénomènes convectifs. Enfin, un algorithme de recalage de modèle éléments finis pondéré sur les données de températures et de déplacements a été implémenté en vue d’identifier un ensemble de propriétés thermiques et thermomécaniques, en tenant compte de la sensibilité de chaque paramètre par rapport aux incertitudes de mesures
The aim of this thesis is firstly to develop procedures of full-field measurements with Digital Image Correlation (DIC), coupled to thermal measurements, suitable for high-temperature experiments on CMC specimens under thermal conditions representative of an engine environment. Secondly, a methodology is proposed for identifying the thermal and thermomechanical properties of the material, quantifying at each stage of the chain the uncertainties associated with the quantities of interest and strategies to reduce them. It was necessary to deal with the challenges due to high temperatures, especially for DIC, either in terms of acquisition (saturation, loss of contrast) or measurement (artefacts due to the mirage effect, also called "heat haze effect").This work has led to the development of a calibration protocol for a multi-instrumented bench using either an in-situ calibration target or by self-calibration using the specimen itself and its environment. 3D surface displacement measurements (with global stereocorrelation approaches) and thermal measurements have made it possible to highlight the heat haze effect phenomenon. Spatiotemporal regularisation strategies of the measured displacements were proposed and allowed satisfactory results to be obtained (significant reduction of measurement uncertainties). Similarly, model reduction approaches (POD) have been used to process thermal data and quantify the uncertainties associated with convective phenomena. Finally, a weighted Finite-Element Model Updating (FEMU) algorithm on both temperature and displacement data was implemented in order to identify a set of thermal and thermomechanical properties, taking into account the sensitivity of each parameter with regard to measurement uncertainties

Cette étude porte sur l’analyse de l’influence de l’eau sur des composites chanvre/époxy. Deux résines ont été utilisées, l’une 100 % pétrochimique (Epolam) et l’autre partiellement bio-sourcée (Greenpoxy). L’absorption d’eau d’éprouvettes de composites tissés a d’abord été suivie par des mesures gravimétriques. Les résultats ont mis en évidence une cinétique d’absorption de type Langmuir dépendant de plusieurs paramètres (types de résine, de fibres et de tissu, fraction volumique de fibres et épaisseur). Des essais de DMA et DSC sur les résines ont montré un effet plastifiant de l’eau ainsi qu’une modification irréversible de leurs propriétés après le vieillissement hydrique. Des observations microscopiques réalisées sur le composite et à l’échelle du fil ont permis de détecter que des fissures entre les fibres de chanvre et la matrice époxy apparaissent et se développent dès le début de la désorption. De surcroît, des observations en photoélasticimétrie ont montré une réorganisation des contraintes internes irréversible due à ces endommagements. Des essais de traction quasi-statiques ont ensuite été réalisés sur des composites tissés et sur les résines seules. Le vieillissement hydrique modifie les grandeurs mécaniques ainsi que l’évolution des endommagements, suivie par émission acoustique et tomographie X.Pour mieux appréhender le comportement de l’interface fil/matrice, différents tests ont été développés sur des éprouvettes monofilamentaires spécifiques dans lesquelles le fil est orienté à 45° ou 90° par rapport à l’axe de sollicitation. Tout d’abord, des tests de nanoindentation ont été effectués de manière à mesurer in situ les propriétés de chaque constituant des composites avant et après vieillissement hydrique. Un second type de test basé sur le principe du choc laser a été utilisé pour déterminer la qualité de l’adhérence entre le fil et la résine. Enfin, des essais de traction sur micromachine sous microscope optique ont été mis au point dans le but de mesurer les champs de déformations par corrélation d’images. Les résultats ont démontré que les déformations locales sont plus importantes après un vieillissement hydrique et que ce phénomène est plus accentué dans le chanvre/Greenpoxy que dans le chanvre/Epolam. Le développement d’un modèle numérique simplifié a permis d’approcher la loi de comportement locale des différents constituants du composite et d’étudier l’évolution des modules après vieillissement
This study deals with the analysis of the influence of water on hemp/epoxy composites. Two resins have been used, one being 100 % petrochemical (Epolam) and the other one partially bio-based (Greenpoxy). The water absorption of the woven composites have been measured with gravimetric analyses. The results showed a Langmuir-type absorption, and depends on several parameters (type of resin, fibres and fabric, fibre ratio and thickness). DMA and DSC tests realised on resins showed the plasticizing effect of water and an irreversible modification of their properties after the water ageing. Microscopic observations of the composite at the yarn scale helped to detect that cracks between hemp fibres and epoxy matrix appear and develop right from the start of the desorption. Moreover, observations by photoelasticimetry showed an irreversible reorganisation of internal stresses due to this damage. Next, quasi-static tensile tests have been performed on woven composites and resins. The water ageing has modified the mechanical properties of the materials as well as the evolution of damage, which has been studied by acoustic emission and X-ray tomography.For a better understanding of the yarn/matrix interface behaviour, different tests have been developed with specific single yarn composites, in which the yarn is oriented at 45° or 90° in regard to the main axis of the specimen. First, nanoindentation tests have been performed, to measure the properties in situ of each component of the composites before and after water ageing. A second type of tests, based on the laser shock wave method, has been used to determine the quality of adhesion between the yarn and the resin. Finally, tensile tests on a micromachine under an optical microscope have been developed with the aim of measuring strain fields by digital image correlation. The results demonstrated that local strains are higher after a water ageing and that this phenomenon is more pronounced in the hemp/Greenpoxy than in the hemp/Epolam. The development of a finite element model led to estimate the local constitutive law of the different components of the composites and to study the evolution of the modulus after the ageing

Graphical models are powerful tools when estimating complex dependence structures among large sets of data. This thesis restricts the scope to undirected Gaussian graphical models. An initial predefined sparse precision matrix was specified to generate multivariate normally distributed data. Utilizing the generated data, a simulation study was conducted reviewing accuracy, sensitivity and specificity of the estimated precision matrix. The graphical LASSO was applied using four different packages available in R with seven selection criteria's for estimating the tuning parameter. The findings are mostly in line with previous research. The graphical LASSO is generally faster and feasible in high dimensions, in contrast to stepwise model selection. A portion of the selection methods for estimating the optimal tuning parameter obtained the true network structure. The results provide an estimate of how well each model obtains the true, predefined dependence structure as featured in our simulation. As the simulated data used in this thesis is merely an approximation of real-world data, one should not take the results as the only aspect of consideration when choosing a model.

It was found that a significant improvement of mechanical properties of CFRPs can be achieved by the adjustment of the matrix properties such as toughness and CF/matrix adhesion via the chemical modification, as well as the physical modification by a small amount of cheap and environment-friendly nano fibers. Based on investigation of fracture mechanisms at macro/micro scale, the effects of matrix properties and nano fiber on the mechanical properties of CFRP have been discussed. Subsequently, the relationship has been characterized by a numerical model to show how to modulate the parameters of the matrix properties to achieve excellent fatigue properties of CFRP.
博士(工学)
Doctor of Philosophy in Engineering
同志社大学
Doshisha University

The purpose of this thesis is to test the Q method applied in discursive analysis in public policy. The author will carry out research using available scientific literature dealing with the method and prepare a case study whose purpose will be to test the reliability of the method Q in a practical case. Promoters of the Q method argue that the uniqueness of the method rests in its resistance to the subjective influence of the researcher. The purpose of this work is therefore a practical verification of to what extent may the researcher's subjective stance at the selection of the default set of statements reflect on the overall result of the analysis.

Dans un contexte où la demande en protéines augmente, il est nécessaire de mieux connaître la qualité nutritionnelle des protéines végétales, et notamment leur digestibilité. Pour cela, les modèles humains et porcins sont préconisés, mais ils sont lourds à mettre en place et posent des questions éthiques. L’objectif de la thèse est donc de mieux comprendre les phénomènes impliqués dans la digestion des protéines végétales et de simuler celle-ci, en cohérence avec la physiologie digestive, à partir de modèles in vitro. Quatre repas avec les mêmes teneurs en macronutriments ont été développés, dont les protéines sont exclusivement apportées par du seitan, du tofu, du jus de soja ou une émulsion de pois. Dans un premier temps, nous avons récolté chez le mini-porc à différents niveaux du segment du système digestif (l’estomac, le duodénum et l’iléon)des données in vivo spécifiques à chaque aliment et nécessaires au paramétrage de systèmes in vitro dynamiques, comme le DiDGI®. Il a été montré que des variables étaient dépendantes de la nature de l’aliment ingéré, comme l’acidification gastrique ou le flux d’arrivée du bol au niveau de l’iléon. Dans un second temps, différentes valeurs relatives à l’état des protéines et notamment leur hydrolyse ont été suivies dans les échantillons prélevés à différents endroits et temps au cours des digestions des différents aliments dans les conditions in vivo et in vitro dynamique. Globalement, de bons résultats de corrélation ont été trouvés entre les deux modèles, montrant que les programmes de digestion aliments-dépendants développés pou
In a context of increasing demand for plant proteins, it is important to properly estimate their nutritional quality, and especially their digestibility. Human and pig models are recommended for that, but these models are complex to implement due to ethical considerations. It is therefore interesting to simulate digestion on in vitro systems. This PhD project aims thus to better understand the phenomena that affect the digestion of plant proteins and to simulate it on in vitro models according to the physiology. Four model meals with the same macronutrient contents have been developed, whose proteins were exclusively provided by the seitan, the tofu, the soya juice or the pea emulsion. First, we collected from the minipig at different levels of their digestive (the stomach, duodenum and the ileum),data for each food necessary to implement dynamic in vitro system (DiDGI®). We observed different kinetics depending on the kind of food. The pH kinetic in the gastric phase and the flux of bolus at the ileal level were specifically impacted. In a second step, different measurements regarding the protein breakdown were followed and compared in the samples collected at different locations and times during the in vivo and the dynamic in vitro food-dependent digestions. Overall, good correlation results were found between the two models, showing that the food-dependent digestion programs developed for DiDGI® can mimic digestion in a consistent manner. It has been shown that some physiological parameters are dependent on the nature of the food ingested, such as the gastric acidifica

Após muitas décadas de mau uso das técnicas construtivas de terra, associando-as à miséria e habitações subumanas, há tentativas de resgatá-las e reintroduzi-las como opção construtiva digna. A terra, utilizada como material construtivo, possui muitas potencialidades, especialmente quando se trata do meio rural, onde se apresenta em quantidade e variedade abundante, possibilitando maior flexibilidade na técnica a ser utilizada. Além disso, as áreas rurais brasileiras, inclusive os assentamentos rurais de reforma agrária, têm um déficit de moradias proporcionalmente equivalente ao das áreas urbanas, porém não recebem uma atenção equivalente na forma de políticas públicas e investimentos. Os programas habitacionais federais, em sua maioria, são desenhados para a população e dinâmica das áreas urbanas e não contemplam particularidades do campo e do trabalhador rural. No Assentamento Rural Sepé Tiaraju, localizado no município de Serra Azul-SP, 77 famílias acessaram recursos provenientes de um programa habitacional federal e do INCRA para construção de suas casas. Dessas 77, um grupo de 10 famílias auto-intitulado de Grupo Alternativo, optou por construir suas casas utilizando materiais ditos mais sustentáveis, não industrializados e, de preferência, não comercializáveis, escolhendo a terra como material construtivo de vedação. No entanto, após cinco anos de projeto, apenas três famílias permaneceram com a terra como opção construtiva. Diante desse contexto, o objetivo dessa pesquisa é identificar e avaliar as variáveis que tiveram maior influência na adoção da terra como material construtivo para habitação social rural no caso do Grupo Alternativo do Assentamento Rural Sepé Tiaraju. Como método de coleta de dados da pesquisa foi utilizado o levantamento documental e áudio-visual, com consulta à literatura técnica e científica sobre o tema, os registros de pesquisa do grupo Habis e consulta a profissionais e pesquisadores da área de construção em terra. Como métodos de análise de dados da pesquisa foram utilizados a Linha do Tempo e a Matriz de Correlação de Influência (MCI). Como resultado, pode-se identificar que as variáveis que tiveram maior influência na adoção da terra como material construtivo para habitação social rural no presente caso, foram as relacionadas ao \"material construtivo\", \"assessor técnico\" e \"produção\". Já as variáveis relacionadas a Famílias, tiveram muita influência, mas não foram determinantes. As relações de influência que tiveram maior destaque foram as relações dentro do grupo de variáveis relacionadas a elementos técnicos e entre essas e as variáveis relacionadas ao assessor técnico. Os resultados dessa pesquisa permitem indicar que a terra é uma opção viável para construção de habitação rural com recursos de programas habitacionais federais, desde que exista um programa habitacional direcionado para as especificidades do meio rural e uso de materiais não convencionais.
After many decades of misuse of earth-based building techniques, which are associated with poverty and subhuman housing, some people still try to rescue and reintroduce them as a worthy alternative. The earth used as a building material has many potential, mainly when applied to rural areas, where this kind of material is encountered in abundant quantity and variety, being possible to apply more techniques. In addition, the Brazilian rural areas, including rural settlements of agrarian reform have a housing deficit that is proportionally equivalent to the urban areas. However, these rural areas do not receive equivalent attention in terms of public policies and investments. Many federal housing programs are designed for the population and dynamic urban areas, thus, sometimes they do not consider the conditions of rural areas and worker particularities. In Sepé Tiaraju Rural Settlement located in the municipality of Serra Azul, state of São Paulo, 77 families used funds obtained from a federal housing program and from INCRA to build their homes. From these 77 families, a group of 10 families chose to build their homes using more sustainable materials, non-industrialized and non-commercialized, choosing the earth as a building material for walls. This group was called Alternative Group. However, after five years only three families remain using the earth-based building technique. Following this context, the objective of this research is the identification and evaluation of variables that most influenced the adoption of earth-based building technique applied to social housing in rural areas. In order to validate the methodology proposed, this work uses the case of the Alternative Group (Sepé Tiaraju Rural Settlement). In this way, as the method to obtain research data was the use of documents and audio-visual resources, the researches of Habis Group, moreover, consultation to professionals and researchers of this area. After obtained all these data, a timeline and a correlation matrix were employed in order to evaluate the data in a qualitative way. It is important to highlight that the correlation matrix was used to verify the influence between the variables that compose the data previously obtained. Hence, one can identify that the variables related to technical advisor, production and building material were most responsible to use the earth-based building technique in this case study. On the other hand, the variables related to family had some influence, but were not decisive, and the variables related to technical elements and their relationships with the variables related to technical advisor have been highlighted. The final results of this research allow us to indicate that earth is a viable choice for building rural housing using funds from federal housing programs. However, it is important that these programs target the specificities of rural areas and unconventional materials.