L'exploration des bases de données temporelles à l'aide de méthodes de fouille de données adaptées a fait l'objet de nombreux travaux de recherche. Cependant le volume d'informations extraites est souvent important et la tâche d'analyse reste alors difficile. Dans cette thèse, nous présentons des méthodes pour synthétiser et filtrer l'information extraite. L'objectif est de restituer des résultats qui soient interprétables. Pour cela, nous avons exploité la notion de séquence partiellement ordonnée et nous proposons (1) un algorithme qui extrait l'ensemble des motifs partiellement ordonnés clos; (2) un post-traitement pour filtrer un ensemble de motifs d'intérêt et(3) une approche qui extrait un consensus comme alternative à l'extraction de motifs. Les méthodes proposées ont été testées sur des données hydrobiologiques issues du projet ANR Fresqueau et elles ont été implantées dans un logiciel de visualisation destiné aux hydrobiologistes pour l'analyse de la qualité des cours d'eau. / Exploring temporal databases with suitable data mining methods have been the subject of several studies. However, it often leads to an excessive volume of extracted information and the analysis is difficult for the user. We addressed this issue and we specically focused on methods that synthesize and filter extracted information. The objective is to provide interpretable results for humans. Thus, we relied on the notion of partially ordered sequence and we proposed (1) an algorithm that extracts the set of closed partially ordered patterns ; (2) a post-processing to filter some interesting patterns for the user and (3) an approach that extracts a partially ordered consensus as an alternative to pattern extraction. The proposed methods were applied for validation on hydrobiological data from the Fresqueau ANR project. In addition, they have been implemented in a visualization tool designed for hydrobiologists for water course quality analysis.

Nesta dissertação apresentamos métodos de diagnóstico em modelos semiparamétricos sob erros normais, em especial os modelos semiparamétricos com uma variável explicativa não paramétrica, conhecidos como modelos lineares parciais. São utilizados splines cúbicos para o ajuste da variável resposta e são aplicadas funções de verossimilhança penalizadas para a obtenção dos estimadores de máxima verossimilhança com os respectivos erros padrão aproximados. São derivadas também as propriedades da matriz hat para esse tipo de modelo, com o objetivo de utilizá-la como ferramenta na análise de diagnóstico. Gráficos normais de probabilidade com envelope gerado também foram adaptados para avaliar a adequabilidade do modelo. Finalmente, são apresentados dois exemplos ilustrativos em que os ajustes são comparados com modelos lineares normais usuais, tanto no contexto do modelo aditivo normal simples como no contexto do modelo linear parcial. / In this master dissertation we present diagnostic methods in semiparametric models under normal errors, specially in semiparametric models with one nonparametric explanatory variable, also known as partial linear model. We use cubic splines for the nonparametric fitting, and penalized likelihood functions are applied for obtaining maximum likelihood estimators with their respective approximate standard errors. The properties of the hat matrix are also derived for this kind of model, aiming to use it as a tool for diagnostic analysis. Normal probability plots with simulated envelope graphs were also adapted to evaluate the model suitability. Finally, two illustrative examples are presented, in which the fits are compared with usual normal linear models, such as simple normal additive and partially linear models.

O objetivo principal desse trabalho é investigar a aplicação de cerâmica á base de Zirconia em ferramentas de corte na usinagem de aço temperado (VND) utilizando um torno de ultraprecisão. Foram analisados dois tipos de composições de cerâmicas com estrutura cristalina diferentes, sendo elas: monoclínica e tetragonal. A diferença destas estruturas é devido à adição de Ytria. A fase monoclínica não contém Ytria em sua composição, enquanto, a fase tetragonal é obtida com Ytria (\'Y IND.3\') (chamada zirconia parcialmente estabilizada com Ytria). A fase tetragonal apresenta uma resistência elevada ao impacto junto com alta dureza (1800 kgf/\'MM POT.2\') quando comparada com a fase monoclínica que apresenta alta dureza mas menor tenacidade. Devido a este fato, esses materiais têm chamado à atenção dos pesquisadores para a usinagem de aços endurecidos. A geração de superfície é influenciada por diversos fatores, sendo eles: material peça, condições de corte, erros macro geométrico, erros de micro geometricos e do estado da aresta da ferramenta. Na usinagem de ultraprecisão a alta rigidez e vibração/trepidação máquina ferramenta é usada para evitar erros de micro geometria e macro geometria que conseqüentemente são transferidos para superfície da peça. Neste trabalho, devido ao fato de se usar um torno de ultraprecisão é possível afirmar que o perfil da rugosidade é gerado pela replicação do perfil da aresta da ferramenta de corte para a superfície da peça. A rugosidade da superfície foi medida com um perfilometro óptico com resolução de 0,1 nm. Os resultados mostraram que a rugosidade da superfície obtida após os testes de usinagem com as ferramantas de cerâmicas chegou á valores em torno de 0,140 microns, o que equivale ao acabamento com processo de retificação. Outro aspecto importante refere-se ao desgaste das ferramentas que, conseqüentemente, tem uma grande influência nos resultados obtidos. As ferramentas de corte foram analisadas antes e depois da usinagem por microscópio eletrônico de varredura. Verificou-se que as ferramentas de corte na fase tetragonal apresentaram desgaste do tipo cratera na aresta da ferramenta enquanto a aresta da ferramenta monoclínica apresentou desgaste do tipo lascamento. / The main objective of this work is to investigate the application of a ceramic composite of Alumina-Zirconia cutting tools inserts in the machining of hardened steel (VND) in an ultraprecision lathe. Two different ceramic compositions with different crystalline structure were tested, to know: monoclinic and tetragonal. The difference in these structures is due the addition of Yttrium. The monoclinic phase has no Yttrium in its composition while the tetragonal phase is obtained with Ytrium (\'Y IND.3\') (named partially stabilized zirconium). The tetragonal phase presents a high impact toughness along with high hardness (1800 kgf/\'MM POT.2\') when compared to the monoclinic phase which presents high hardness but lower toughness. Due to this fact, these materials have draw attention of researchers in the field of machining of hardened steels. The surface generation is influenced by several factors, to know: workpiece material, cutting conditions, macro geometry errors, micro geometry errors and the sharpness of the cutting edge. In ultraprecision machining, a high stiffness and chatter/vibration free machine tool is used in order to avoid common macro and micro geometry errors replicated into the workpiece surface. In this case, it is possible to assert that the roughness profile is generated by the replication of the cutting tool edge profile to the workpiece surface. The surface roughness was measured by an optical profiler with resolution of 0,1 nm. The results showed that the surface roughness obtained after machining tests with these ceramic inserts were in the range of 0,140 micrometers, which is in the same range of roughness obtained by the grinding process. Another important aspect refers to the wear of the ceramic inserts which has direct influence in the performance as a cutting tool material. The cutting inserts were evaluated before and after machining by scanning electron microscope. It was found that the tetragonal phase cutting tools presented crater wear on the rake face while the monoclinic phase presented cutting edge chipping as the main main type of wear.

Ce mémoire comporte deux parties distinctes. La première partie concerne une étude d'algèbres n-aires. Une algèbre n-aire est un espace vectoriel sur lequel est définie une multiplication sur n arguments. Classiquement les multiplications sont binaires, mais depuis l'utilisation en physique théorique de multiplications ternaires, comme les produits de Nambu, de nombreux travaux mathématiques se sont focalisés sur ce type d'algèbres. Deux classes d'algèbres n-aires sont essentielles: les algèbres n-aires associatives et les algèbres n-aires de Lie. Nous nous intéressons aux deux classes. Concernant les algèbres n-aires associatives, on s'intéresse surtout aux algèbres 3-aires partiellement associatives, c'est-à-dire dont la multiplication vérifie l'identité ((xyz)tu)+(x(yzt)u)+(xy(ztu))=0 Ce cas est intéressant car les travaux connus concernant ce type d'algèbres ne distinguent pas les cas n pair et n-impair. On montre dans cette thèse que le cas n=3 ne peut pas être traité comme si n était pair. On étudie en détail l'algèbre libre 3-aire partiellement associative sur un espace vectoriel de dimension finie. Cette algèbre est graduée et on calcule précisément les dimensions des 7 premières composantes. On donne dans le cas général un système de générateurs ayant la propriété qu'une base est donnée par la sous famille des éléments non nuls. Les principales conséquences sont L'algèbre libre 3-aire partiellement associative est résoluble. L'algèbre libre commutative 3-aire partiellement associative est telle que tout produit concernant 9 éléments est nul. L'opérade quadratique correspondant aux algèbres 3-aires partiellement associatives ne vérifient pas la propriété de Koszul. On s'intéresse ensuite à l'étude des produits n-aires sur les tenseurs. L'exemple le plus simple est celui d'un produit interne sur des matrices non carrées. Nous pouvons définir le produit 3aire donné par A . ^tB . C. On montre qu'il est nécessaire de généraliser un peu la définition de partielle associativité. Nous introduisons donc les produits -partiellement associatifs où  est une permutation de degré p. Concernant les algèbres de Lie n-aires, deux classes d'algèbres ont été définies: les algèbres de Fillipov (aussi appelées depuis peu les algèbres de Lie-Nambu) et les algèbres n-Lie. Cette dernière notion est très générale. Cette dernière notion, très important dans l'étude de la mécanique de Nambu-Poisson, est un cas particulier de la première. Mais pour définir une approche du type Maurer-Cartan, c'est-à-dire définir une cohomologie scalaire, nous considérons dans ce travail les algèbres de Fillipov comme des algèbres n-Lie et développons un tel calcul dans le cadre des algèbres n-Lie. On s'intéresse également à la classification des algèbres n-aires nilpotentes. Le dernier chapitre de cette partie est un peu à part et reflète un travail poursuivant mon mémoire de Master. Il concerne les algèbres de Poisson sur l'algèbre des polynômes. On commence à présenter le crochet de Poisson sous forme duale en utilisant des équations de Pfaff. On utilise cette approche pour classer les structures de Poisson non homogènes sur l’algèbre des polynômes à trois variables . Le lien avec les algèbres de Lie est clair. Du coup on étend notre étude aux algèbres de Poisson dont l'algèbre de Lie sous jacent est rigide et on applique les résultats aux algèbres enveloppantes des algèbres de Lie rigides. La partie 2 concerne l'arithmétique des intervalles. Cette étude a été faite suite à une rencontre avec une société d'ingénierie travaillant sur des problèmes de contrôle de paramètres, de problème inverse (dans quels domaines doivent évoluer les paramètres d'un robot pour que le robot ait un comportement défini). [...] / This thesis has two distinguish parts. The first part concerns the study of n-ary algebras. A n-ary algebra is a vector space with a multiplication on n arguments. Classically the multiplications are binary, but the use of ternary multiplication in theoretical physic like for Nambu brackets led mathematicians to investigate these type of algebras. Two classes of n-ary algebras are fundamental: the associative n-ary algebras and the Lie n-ary algebras. We are interested by both classes. Concerning the associative n-ary algebras we are mostly interested in 3-ary partially associative 3-ary algebras, that is, algebras whose multiplication satisfies ((xyz)tu)+(x(yzt)u)+(xy(ztu))=0. This type is interesting because the previous woks on this subject was not distinguish the even and odd cases. We show in this thesis that the case n=3 can not be treated as the even cases. We investigate in detail the free partially associative 3-ary algebra on k generators. This algebra is graded and we compute the dimensions of the 7 first components. In the general case, we give a spanning set such as the sub family of non zero vector is a basis. The main consequences are the free partially associative 3-ary algebra is solvable. In the free commutative partially associative 3-ary algebra any product on 9 elements is trivial. The operad for partially associative 3-ary algebra do not satisfy the Koszul property. Then we study n-ary products on the tensors. The simplest example is given by a internal product of non square matrices. We can define a 3-ary product by taking A . ^tB . C. We show that we have to generalize a bit the definition of partial associativity for n-ary algebras. We then introduce the products -partially associative where  is a permutation of the symmetric group of degree n. Concerning the n-ary algebras, two classes have been defined: Filipov algebras (also called recently Lie-Nambu algebras) and some more general class, the n-Lie algebras. Filipov algebras are very important in the study of the mechanic of Nambu-Poisson, and is a particular case of the other. So to define an approach of Maurer-Cartan type, that is, define a scalar cohomology, we consider in this work Fillipov as n-Lie algebras and develop such a calculus in the n-Lie algebras frame work. We also give some classifications of n-ary nilpotent algebras. The last chapter of this part concerns my work in Master on the Poisson algebras on polynomials. We present link with the Lie algebras is clear. Thus we extend our study to Poisson algebras which associated Lie algebra is rigid and we apply these results to the enveloping algebras of rigid Lie algebras. The second part concerns intervals arithmetic. The interval arithmetic is used in a lot of problems concerning robotic, localization of parameters, and sensibility of inputs. The classical operations of intervals are based of the rule : the result of an operation of interval is the minimal interval containing all the result of this operation on the real elements of the concerned intervals. But these operations imply many problems because the product is not distributive with respect the addition. In particular it is very difficult to translate in the set of intervals an algebraic functions of a real variable. We propose here an original model based on an embedding of the set of intervals on an associative algebra. Working in this algebra, it is easy to see that the problem of non distributivity disappears, and the problem of transferring real function in the set of intervals becomes natural. As application, we study matrices of intervals and we solve the problem of reduction of intervals matrices (diagonalization, eigenvalues, and eigenvectors).

In HIV/AIDS studies, viral load (the number of copies of HIV-1 RNA) and CD4 cell counts are important biomarkers of the severity of viral infection, disease progression, and treatment evaluation. Recently, joint models, which have the capability on the bias reduction and estimates' efficiency improvement, have been developed to assess the longitudinal process, survival process, and the relationship between them simultaneously. However, the majority of the joint models are based on mean regression, which concentrates only on the mean effect of outcome variable conditional on certain covariates. In fact, in HIV/AIDS research, the mean effect may not always be of interest. Additionally, if obvious outliers or heavy tails exist, mean regression model may lead to non-robust results. Moreover, due to some data features, like left-censoring caused by the limit of detection (LOD), covariates with measurement errors and skewness, analysis of such complicated longitudinal and survival data still poses many challenges. Ignoring these data features may result in biased inference. Compared to the mean regression model, quantile regression (QR) model belongs to a robust model family, which can give a full scan of covariate effect at different quantiles of the response, and may be more robust to extreme values. Also, QR is more flexible, since the distribution of the outcome does not need to be strictly specified as certain parametric assumptions. These advantages make QR be receiving increasing attention in diverse areas. To the best of our knowledge, few study focuses on the QR-based joint models and applies to longitudinal-survival data with multiple features. Thus, in this dissertation research, we firstly developed three QR-based joint models via Bayesian inferential approach, including: (i) QR-based nonlinear mixed-effects joint models for longitudinal-survival data with multiple features; (ii) QR-based partially linear mixed-effects joint models for longitudinal data with multiple features; (iii) QR-based partially linear mixed-effects joint models for longitudinal-survival data with multiple features. The proposed joint models are applied to analyze the Multicenter AIDS Cohort Study (MACS) data. Simulation studies are also implemented to assess the performance of the proposed methods under different scenarios. Although this is a biostatistical methodology study, some interesting clinical findings are also discovered.

In this paper we present a component based person detection system that is capable of detecting frontal, rear and near side views of people, and partially occluded persons in cluttered scenes. The framework that is described here for people is easily applied to other objects as well. The motivation for developing a component based approach is two fold: first, to enhance the performance of person detection systems on frontal and rear views of people and second, to develop a framework that directly addresses the problem of detecting people who are partially occluded or whose body parts blend in with the background. The data classification is handled by several support vector machine classifiers arranged in two layers. This architecture is known as Adaptive Combination of Classifiers (ACC). The system performs very well and is capable of detecting people even when all components of a person are not found. The performance of the system is significantly better than a full body person detector designed along similar lines. This suggests that the improved performance is due to the components based approach and the ACC data classification structure.

The focus of this study is the development and evaluation of a new partially supervised learning framework. This framework belongs to an emerging field in machine learning that augments unsupervised learning processes with some elements of supervision. It is based on proximity fuzzy clustering, where an active learning process is designed to query for the domain knowledge required in the supervision. Furthermore, the framework is extended to the parametric optimization of the kernel function in the proximity fuzzy clustering algorithm, where the goal is to achieve interesting non-spherical cluster structures through a non-linear mapping. It is demonstrated that the performance of kernel-based clustering is sensitive to the selection of these kernel parameters. Proximity hints procured from domain knowledge are exploited in the partially supervised framework. The theoretic developments with proximity fuzzy clustering are evaluated in several interesting and practical applications. One such problem is the clustering of a set of graphs based on their structural and semantic similarity. The segmentation of music is a second problem for proximity fuzzy clustering, where the aim is to determine the points in time, i.e. boundaries, of significant structural changes in the music. Finally, a time series prediction problem using a fuzzy rule-based system is established and evaluated. The antecedents of the rules are constructed by clustering the time series using proximity information in order to localize the behavior of the rule consequents in the architecture. Evaluation of these efforts on both synthetic and real-world data demonstrate that proximity fuzzy clustering is well suited for a variety of problems. / Digital Signals and Image Processing

In an experiment to understand colon carcinogenesis, all animals were exposed to a carcinogen while half the animals were also exposed to radiation. Spatially, we measured the existence of aberrant crypt foci (ACF), namely morphologically changed colonic crypts that are known to be precursors of colon cancer development. The biological question of interest is whether the locations of these ACFs are spatially correlated: if so, this indicates that damage to the colon due to carcinogens and radiation is localized. Statistically, the data take the form of binary outcomes (corresponding to the existence of an ACF) on a regular grid. We develop score??type methods based upon the Matern and conditionally autoregression (CAR) correlation models to test for the spatial correlation in such data, while allowing for nonstationarity. Because of a technical peculiarity of the score??type test, we also develop robust versions of the method. The methods are compared to a generalization of Moran??s test for continuous outcomes, and are shown via simulation to have the potential for increased power. When applied to our data, the methods indicate the existence of spatial correlation, and hence indicate localization of damage. Assuming that there are correlations in the locations of the ACF, the questions are how great are these correlations, and whether the correlation structures di?er when an animal is exposed to radiation. To understand the extent of the correlation, we cast the problem as a spatial binary regression, where binary responses arise from an underlying Gaussian latent process. We model these marginal probabilities of ACF semiparametrically, using ?xed-knot penalized regression splines and single-index models. We ?t the models using pairwise pseudolikelihood methods. Assuming that the underlying latent process is strongly mixing, known to be the case for many Gaussian processes, we prove asymptotic normality of the methods. The penalized regression splines have penalty parameters that must converge to zero asymptotically: we derive rates for these parameters that do and do not lead to an asymptotic bias, and we derive the optimal rate of convergence for them. Finally, we apply the methods to the data from our experiment.

Three different topics of turbulence research that cover modeling, theory and model computation categories are selected and studied in depth. In the first topic, "velocity gradient dynamics in turbulence" (modeling), the Lagrangian linear diffusion model that accounts for the viscous-effect is proposed to make the existing restricted-Euler velocity gradient dynamics model quantitatively useful. Results show good agreement with DNS data. In the second topic, "pressure-strain correlation in homogeneous anisotropic turbulence subject to rapid strain-dominated distortion" (theory), extensive rapid distortion calculation is performed for various anisotropic initial turbulence conditions in strain-dominated mean flows. The behavior of the rapid pressure-strain correlation is investigated and constraining criteria for the rapid pressure-strain correlation models are developed. In the last topic, "unsteady computation of turbulent flow past a square cylinder using partially-averaged Navier-Stokes method" (model computation), the basic philosophy of the PANS method is reviewed and a practical problem of flow past a square cylinder is computed for various levels of physical resolution. It is revealed that the PANS method can capture many important unsteady flow features at an affordable computational effort.

Planning in partially observable domains is a notoriously difficult problem. However, in many real-world scenarios, planning can be simplified by decomposing the task into a hierarchy of smaller planning problems which, can then be solved independently of one another. Several approaches, mainly dealing with fully observable domains, have been proposed to optimize a plan that decomposes according to a hierarchy specified a priori. Some researchers have also proposed to discover hierarchies in fully observable domains. In this thesis, we investigate the problem of automatically discovering planning hierarchies in partially observable domains. The main advantage of discovering hierarchies is that, for a plan of a fixed size, hierarchical plans can be more expressive than non-hierarchical ones. Our solution frames the discovery and optimization of a hierarchical policy as a non-convex optimization problem. By encoding the hierarchical structure as variables of the optimization problem, we can automatically discover a hierarchy. Successfully solving the optimization problem therefore yields an optimal hierarchy and an optimal policy. We describe several techniques to solve the optimization problem. Namely, we provide results using general non-linear solvers, mixed-integer linear and non-linear solvers or a form of bounded hierarchical policy iteration. Our method is flexible enough to allow any parts of the hierarchy to be specified based on prior knowledge while letting the optimization discover the unknown parts. It can also discover hierarchical policies, including recursive policies, that are more compact (potentially infinitely fewer parameters).