Dissertations / Theses on the topic 'Modèle statistique linéaire'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Modèle statistique linéaire.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Dauxois, Thierry. "Dynamique non linéaire et mécanique statistique d'un modèle d'ADN." Dijon, 1993. http://www.theses.fr/1993DIJOS003.
Full textLyazrhi, Faouzi. "Procédures optimales de détection de ruptures dans un modèle linéaire gaussien." Toulouse 3, 1993. http://www.theses.fr/1993TOU30076.
Full textPiet-Lahanier, Hélène. "Estimation de paramètres pour des modèles à erreur bornée." Paris 11, 1987. http://www.theses.fr/1987PA112203.
Full textKhraibani, Hussein. "Modélisation statistique de données longitudinales sur un réseau routier entretenu." Ecole centrale de Nantes, 2010. http://www.theses.fr/2010ECDN0040.
Full textRoad transportation has a direct impact on a country's economy. Infrastructures, particularly pavements, deteriorate under the effect of traffic and climate. As a result, they most constantly undergo maintenance which often requires expensive works. The optimization of maintenance strategies and the scheduling of works necessarily pass by a study that makes use of deterioration evolution laws and accounts for the effect of maintenance on these laws. In this respect, numerous theoretical and experimental works ranging linear and nonlinear regressions to more sophisticated methods such as Markov chain have been conducted. The thesis presents a survey of models and methods and focuses on the analysis of survival data (MADS), an analysis which constituted the objective of important works at LCPC. In order to acount for the fact that current databases contain repeated measurements of each pavement section, the thesis proposes a different approach based on the use of nonlinear mixed-effects models (NLME). First, it carries out a comparison between the NLME and MADS models on different databases in terms of the goodness of fit and prediction capability. The comparison then allows to draw conclusions about the applicability of the two models
Seck, Ousmane. "Sur un modèle de diffusion non linéaire en dynamique des populations." Nancy 1, 1986. http://www.theses.fr/1986NAN10162.
Full textSaumard, Mathieu. "Contribution à l'analyse statistique des données fontionnelles." Thesis, Rennes, INSA, 2013. http://www.theses.fr/2013ISAR0009/document.
Full textIn this thesis, we are interested in the functional data. The problem of estimation in a model of estimating equations is studying. We derive a central limit type theorem for the considered estimator. The optimal instruments are estimated, and we obtain a uniform convergence of the estimators. We are then interested in various testing with functional data. We study the problem of nonparametric testing for the effect of a random functional covariate on an error term which could be directly observed as a response or estimated from a functional model like for instance the functional linear model. We proved, in order to construct the tests, a result of dimension reduction which relies on projections of the functional covariate. We have constructed no-effect tests by using a kernel smoothing or a nearest neighbor smoothing. A goodness-of-fit test in the functional linear model is also proposed. All these tests are studied from a theoretical and practical perspective
Roget-Vial, Céline. "deux contributions à l'étude semi-paramétrique d'un modèle de régression." Phd thesis, Université Rennes 1, 2003. http://tel.archives-ouvertes.fr/tel-00008730.
Full textBoisbunon, Aurélie. "Sélection de modèle : une approche décisionnelle." Phd thesis, Université de Rouen, 2013. http://tel.archives-ouvertes.fr/tel-00793898.
Full textRohart, Florian. "Prédiction phénotypique et sélection de variables en grande dimension dans les modèles linéaires et linéaires mixtes." Thesis, Toulouse, INSA, 2012. http://www.theses.fr/2012ISAT0027/document.
Full textRecent technologies have provided scientists with genomics and post-genomics high-dimensional data; there are always more variables that are measured than the number of individuals. These high dimensional datasets usually need additional assumptions in order to be analyzed, such as a sparsity condition which means that only a small subset of the variables are supposed to be relevant. In this high-dimensional context we worked on a real dataset which comes from the pig species and high-throughput biotechnologies. Metabolomic data has been measured with NMR spectroscopy and phenotypic data has been mainly obtained post-mortem. There are two objectives. On one hand, we aim at obtaining good prediction for the production phenotypes and on the other hand we want to pinpoint metabolomic data that explain the phenotype under study. Thanks to the Lasso method applied in a linear model, we show that metabolomic data has a real prediction power for some important phenotypes for livestock production, such as a lean meat percentage and the daily food consumption. The second objective is a problem of variable selection. Classic statistical tools such as the Lasso method or the FDR procedure are investigated and new powerful methods are developed. We propose a variable selection method based on multiple hypotheses testing. This procedure is designed to perform in linear models and non asymptotic results are given under a condition on the signal. Since supplemental data are available on the real dataset such as the batch or the family relationships between the animals, linear mixed models are considered. A new algorithm for fixed effects selection is developed, and this algorithm turned out to be faster than the usual ones. Thanks to its structure, it can be combined with any variable selection methods built for linear models. However, the convergence property of this algorithm depends on the method that is used. The multiple hypotheses testing procedure shows good empirical results. All the mentioned methods are applied to the real data and biological relationships are emphasized
Guégan, Dominique. "Modèles bilinéaires et polynomiaux de séries chronologiques : étude probabiliste et analyse statistique." Grenoble 1, 1988. http://tel.archives-ouvertes.fr/tel-00330671.
Full textGasseling, Tony. "Caractérisation non linéaire avancée de transistors de puissance pour la validation de leur modèle CAO." Limoges, 2003. http://www.theses.fr/2003LIMO0041.
Full textAdvanced functional characterizations of power transistors for the validation of nonlinear models of SC devices used in CAD packages. This work deals with different functional characterization methods for the design of optimized power amplifiers. These characterizations are carried out on transistors at the first stages of the design, in a source and load-pull environment. Thus, it is shown that a pulsed load-pull set up is very useful to validate the technologies used for the generation of high power at RF and microwave frequencies. It also enables to deeply validate the thermoelectric nonlinear models of transistors developed for this purpose. For the design of amplifiers which operate up to millimetric frequencies (Ku / K Band), reaching high power under constraint of efficiency and linearity is one of the most critical point because of the weak reserves of power gain proposed. In this context, the development of an active source and load-pull setup is of prime importance. It enables to primarily determine the transistor optimum operating conditions (Matching and DC bias) to reach the best trade off between efficiency and linearity. Finally, a new method to perform Hot Small-Signal S-Parameter measurements of power transistors operating under large signal conditions is proposed. An application to the prediction of parametric oscillations when the transistor is driven by a pump signal is demonstrated
Kapche, Tagne François. "Utilisation d'un modèle quasi-bidimensionnel pour la simulation d'un transistor à effet de champ en régime de fonctionnement non linéaire." Lille 1, 1997. http://www.theses.fr/1997LIL10037.
Full textPicard, Guillaume. "Traitement statistique des distorsions non-linéaires pour la restauration des enregistrements sonores." Phd thesis, Télécom ParisTech, 2006. http://pastel.archives-ouvertes.fr/pastel-00002315.
Full textRibereau, Pierre. "Quelques contributions à la statistique théorique et appliquée." Paris 6, 2005. http://www.theses.fr/2005PA066164.
Full textCrambes, Christophe. "Modèles de régression linéaire pour variables explicatives fonctionnelles." Phd thesis, Université Paul Sabatier - Toulouse III, 2006. http://tel.archives-ouvertes.fr/tel-00134003.
Full textThébaud, Olivier. "Emploi des chaînes de Markov dérivantes dans l'étude du génome." Paris 5, 2001. http://www.theses.fr/2001PA05S008.
Full textLaurent-Chabalier, Sabine. "Impact d'une mauvaise spécification de la variance sur la statistique du test F d'un modèle linéaire : étude de séries temporelles de richesse en sucre de la canne." Montpellier 2, 2007. http://www.theses.fr/2007MON20192.
Full textHernandez, Quintero Angelica. "Inférence statistique basée sur les processus empiriques dans des modèles semi-paramétriques de durées de vie." Toulouse 3, 2010. http://thesesups.ups-tlse.fr/1201/.
Full textSurvival data arise from disciplines such as medicine, criminology, finance and engineering amongst others. In many circumstances the event of interest can be classified in several causes of death or failure and in some others the event can only be observed for a proportion of "susceptibles". Data for these two cases are known as competing risks and long-term survivors, respectively. Issues relevant to the analysis of these two types of data include basic properties such as the parameters estimation, existence, consistency and asymptotic normality of the estimators, and their efficiency when they follow a semiparametric structure. The present thesis investigates these properties in well established semiparametric formulations for the analysis of both competing risks and long-term survivors. It presents an overview of mathematical tools that allow for the study of these basic properties and describes how the modern theory of empirical processes and the theory of semiparametric efficiency facilitate relevant proofs. Also, consistent variance estimate for both the parametric and semiparametric components for the two models are presented. The findings of this research provide the theoretical basis for obtaining inferences with large samples, the calculation of confidence bands and hypothesis testing. The methods are illustrated with data bases generated through simulations
Cellier, Dominique. "Deux extensions des résultats classiques sur les estimateurs à rétrécisseur : cas de rétrécisseurs non différentiables ; cas de lois à symétrie sphérique." Rouen, 1986. http://www.theses.fr/1986ROUES044.
Full textKumar, Vandhna. "Descente d'échelle statistique du niveau de la mer pour les îles du Pacifique Sud-Ouest : une approche de régression linéaire multiple." Thesis, Toulouse 3, 2019. http://www.theses.fr/2019TOU30234.
Full textSea level rise is a growing concern in the islands of the western Pacific. Over the altimetry era (1993-present), sea level rise rates in the western tropical Pacific were amongst the highest recorded across the world ocean, reaching up to 3-4 times the global mean. As more and more affected communities relocate to higher grounds to escape the rising seas, there is a compelling need for information on local scales to ease the adaptation and planning process. This is not a straightforward process as sea level varies regionally, driven by wind and ocean circulation patterns, and the prevailing climate modes (e.g. ENSO, PDO/IPO). On local scales, substantial sea level changes can result from natural or anthropogenic induced vertical ground motion. Motivated by such concerns, this thesis focuses on developing a statistical downscaling technique, namely a multiple linear regression (MLR) model, to simulate island sea levels at selected sites in the southwest Pacific - Suva and Lautoka in Fiji, and Nouméa in New Caledonia. The model is based on the knowledge that sea level variations in the tropical Pacific are mainly thermosteric in nature (temperature-related changes in ocean water density) and that these thermosteric variations are dominated by wind-forced, westward propagating Rossby waves. The MLR experiments are conducted over the 1988-2014 study period, with a focus on interannual-to-decadal sea level variability and trend. Island sea levels are first expressed a sum of steric and mass changes. Then, a more dynamical approach using wind stress curl as a proxy for the thermosteric component is undertaken to construct the MLR model. In the latter case, island sea levels are perceived as a composite of global, regional and local components, where the second is dominant. The MLR model takes wind stress curl as the dominant regional regressor (via a Rossby wave model), and the local halosteric component (salinity-related changes in ocean water density), local wind stress, and local sea surface temperature as minor regressors. A stepwise regression function is used to isolate statistically significant regressors before calibrating the MLR model. The modeled sea level shows high agreement with observations, capturing 80% of the variance on average. Stationarity tests on the MLR model indicate that it can be applied skillfully to projections of future sea level. The statistical downscaling approach overall provides insights on key drivers of sea level variability at the selected sites, showing that while local dynamics and the global signal modulate sea level to a given extent, most of the variance is driven by regional factors. [...]
Cogranne, Rémi. "Détection statistique d'informations cachées dans une image naturelle à partir d'un modèle physique." Phd thesis, Université de Technologie de Troyes, 2012. http://tel.archives-ouvertes.fr/tel-00706171.
Full textCaron, Emmanuel. "Comportement des estimateurs des moindres carrés du modèle linéaire dans un contexte dépendant : Étude asymptotique, implémentation, exemples." Thesis, Ecole centrale de Nantes, 2019. http://www.theses.fr/2019ECDN0036.
Full textIn this thesis, we consider the usual linear regression model in the case where the error process is assumed strictly stationary.We use a result from Hannan (1973) who proved a Central Limit Theorem for the usual least squares estimator under general conditions on the design and on the error process. Whatever the design and the error process satisfying Hannan’s conditions, we define an estimator of the asymptotic covariance matrix of the least squares estimator and we prove its consistency under very mild conditions. Then we show how to modify the usual tests on the parameter of the linear model in this dependent context. We propose various methods to estimate the covariance matrix in order to correct the type I error rate of the tests. The R package slm that we have developed contains all of these statistical methods. The procedures are evaluated through different sets of simulations and two particular examples of datasets are studied. Finally, in the last chapter, we propose a non-parametric method by penalization to estimate the regression function in the case where the errors are Gaussian and correlated
Wisniewski, Teodor. "Modélisation non-linéaire des machines synchrones pour l'analyse en régimes transitoires et les études de stabilité." Thesis, Université Paris-Saclay (ComUE), 2018. http://www.theses.fr/2018SACLC092/document.
Full textThe research presented in this thesiswas carried out in the research and developmentproject between Leroy Somer and the Group ofElectrical Engineering of Paris (GeePs). Theirobjective is to simulate the phenomena observedin the transient states of electrical machines.These simulations are particularly oriented bythe new Grid Code requirements for alternatorsconnected to the power network. Two types ofmodels have been principally developed. Thefirst one is based on a magnetic description ofthe machine where each flux is expressed as afunction of the currents flowing through thedifferent machine windings. The second oneregroups the different winding currents by usingthe magnetizing currents on axes d and qassociated to saturation coefficients for eachflux linkage and simplifies the magneticdescription, especially when taking into accountthe damper windings. With a sufficiently precisemodelling of the non-linear magnetic behaviourof the machine, it is possible to better predict thecurrents and the electromagnetic torque underfault conditions such as voltage drops. The workcarried out in this thesis has made possible,starting from the descriptions of the saturationeffects found in a machine, to define methodsfor incorporating saturation into circuit models.Finally, one can make a choice of the dynamicnon-linear model for a given machine. Thanks toshort computation time, it also led to theSimulink integration of the machine andexcitation system models paving the way forstability and control studies
Robert, Christian P. "Résultats nouveaux sur les estimateurs à rétrecisseurs scalaires et matriciels." Rouen, 1987. http://www.theses.fr/1987ROUES029.
Full textBernadet, Pascal. "Contribution à une étude de la prise de décision en sport : approche par la psychologie différentielle et esquisse d'un modèle morphodynamique en activités physiques d'environnement." Bordeaux 2, 2003. http://www.theses.fr/2003BOR21023.
Full textThe present reflexion has for ambition to investigate the mysteries of the decision-taking in sports environment. The basic hypothesis postulates that this phenomenon can be analyzed as a no-linear dynamic system which structures the actor's style. Articulated, within the framework of the Physical Activities of Environment, around the personality of subject, its autobiographical and cultural history and the "wild thought", the decision-making style is in line with a bimodality of confrontation or coalescence with the physical elements of natural environment. The expertise of the subject and the complexity of the situation in which the subject is, can play the role of parameters of control in the emergence of a decision-making dynamics considered under the angle of models of the Catastrophe theory. The analysis which is based on the mathematical theory of singularity tries to model the decision-making functioning of the subject by varying these parameters of control. The various hypothesis checking uses different processes of psychology (parametrical tests and projective tests, interviews) as well as a perspective qualified of morphodynamic. This orientation consists from an observation in real situation, of the motor behavior in two sports practices, the surfboard and the kayak, made contact with the clinical analysis, to deduce the functioning of the psychological substratum able to engender them. The results show the validity of the concept of decision-making style and the elements which constructs it, as well as, on the paradigmatical plan, that of the model of cups type. Finally, the study shows the fertility of a connaction of a differential approach in psychology, a mathematical theory of the singularity and the phenomenological perspective of the decision-making lived, within the framework of a comprehensive analyse of a sports subject in situation
Bayle, Severine. "Modélisation statistique de données fonctionnelles environnementales : application à l'analyse de profils océanographiques." Thesis, Aix-Marseille, 2014. http://www.theses.fr/2014AIXM4016.
Full textTo study biogeochemical processes in the Southern Ocean, tags placed on elephant seals allowed to collect during 2009-2010 oceanographic variables profiles (Chlorophyll a (Chl a), temperature, salinity, light) in an area ranging from southern Kerguelen until the Antarctic continent. This thesis focuses on Chl a data as it is contained in photosynthetic organisms and these ones play an essential role in the oceanic carbon cycle. The infrequently collected vertical Chl a profiles don't provide a mapping of this variable in this area of the ocean. However, we have light profiles sampled more often. The aim of this thesis was then to develop a methodology for reconstructing indirectly Chl a profiles from light profiles, and that takes into account characteristics of this kind of data that naturally occur as functional data. For this, we adressed the profiles decomposition to rebuild or explanations on splines basis, as well as issues related adjustment. A functional linear model was used to predict Chl a profiles from light profiles derivatives. It was shown that the use of such a model provides a good quality of reconstruction to access high frequency variations of Chl a profiles at fine scale. Finally, a functional kriging interpolation predicted the Chl a concentration during night, as light measurements acquired at that time can't be exploited. In the future, the methodology aims to be applied to any type of functional data
Harti, Mostafa. "Estimation robuste sous un modèle de contamination non symétrique et M-estimateur multidimensionnel." Nancy 1, 1986. http://www.theses.fr/1986NAN10063.
Full textSaad, Mohamad. "Méthodes statistiques et stratégies d'études d'association de phénotypes complexes : études pan-génomiques de la maladie de Parkinson." Toulouse 3, 2012. http://thesesups.ups-tlse.fr/1657/.
Full textMy thesis has focused on statistical methods and strategies to study the genetic components of complex human traits and especially of Parkinson's Disease (PD). My work was developed mainly in two contexts of genome wide association studies (GWAS): the detection of common variants and the detection of rare variants. GWAS is an optimal approach in which we have to control for the type I error and the type II error rates. Indeed, a large number of tests are performed. In addition, we must control for potential population stratification problems. Despite the large sample sizes in recent GWASs based on the single-marker test, they may have individually low power to detect common variants with small effects. The use of the multi-marker test may optimize the coverage of genetic variability and thus increase the power of GWAS. I have focused on the study of these tests, especially the "SNP-Set" test based on kernel machine regression and the haplotypic test. I studied the theoretical aspects of these tests and I evaluated the statistical properties in our empirical data for PD. In addition, in our analyses for PD, I developed imputation and meta-analysis techniques to increase the coverage of the genetic variability and the sample size. Association analysis for rare variants faces several challenges. The single marker test is not powerful to detect such variants and the cost of whole-genome sequence analyses for complex traits is still prohibitive. Our design is a cost-effective alternative which is based on the joint use of public sequence data and GWAS data. Several new tests have been proposed but, to date, their statistical properties are still unclear. On the genome-wide level, the type I error and the type II error rates may depend on several factors as gene length, allelic heterogeneity in the gene, LD between SNPs, overlap between genes and the correlation between the common variants and the trait. I evaluated the statistical properties of several methods in simulated data and also in our GWAS PD data. We show that several methods, based on the linear mixed model, are mathematically equivalent and some are special cases of others. In conclusion, we developed strategies and analytical methods which combine complementary approaches (Common Disease-Common Variant versus Common Disease-Rare Variant) to optimize the characterization of the genetic components of PD in particular and of complex traits in general
Perthame, Emeline. "Stabilité de la sélection de variables pour la régression et la classification de données corrélées en grande dimension." Thesis, Rennes 1, 2015. http://www.theses.fr/2015REN1S122/document.
Full textThe analysis of high throughput data has renewed the statistical methodology for feature selection. Such data are both characterized by their high dimension and their heterogeneity, as the true signal and several confusing factors are often observed at the same time. In such a framework, the usual statistical approaches are questioned and can lead to misleading decisions as they are initially designed under independence assumption among variables. The goal of this thesis is to contribute to the improvement of variable selection methods in regression and supervised classification issues, by accounting for the dependence between selection statistics. All the methods proposed in this thesis are based on a factor model of covariates, which assumes that variables are conditionally independent given a vector of latent variables. A part of this thesis focuses on the analysis of event-related potentials data (ERP). ERPs are now widely collected in psychological research to determine the time courses of mental events. In the significant analysis of the relationships between event-related potentials and experimental covariates, the psychological signal is often both rare, since it only occurs on short intervals and weak, regarding the huge between-subject variability of ERP curves. Indeed, this data is characterized by a temporal dependence pattern both strong and complex. Moreover, studying the effect of experimental condition on brain activity for each instant is a multiple testing issue. We propose to decorrelate the test statistics by a joint modeling of the signal and time-dependence among test statistics from a prior knowledge of time points during which the signal is null. Second, an extension of decorrelation methods is proposed in order to handle a variable selection issue in the linear supervised classification models framework. The contribution of factor model assumption in the general framework of Linear Discriminant Analysis is studied. It is shown that the optimal linear classification rule conditionally to these factors is more efficient than the non-conditional rule. Next, an Expectation-Maximization algorithm for the estimation of the model parameters is proposed. This method of data decorrelation is compatible with a prediction purpose. At last, the issues of detection and identification of a signal when features are dependent are addressed more analytically. We focus on the Higher Criticism (HC) procedure, defined under the assumptions of a sparse signal of low amplitude and independence among tests. It is shown in the literature that this method reaches theoretical bounds of detection. Properties of HC under dependence are studied and the bounds of detectability and estimability are extended to arbitrarily complex situations of dependence. Finally, in the context of signal identification, an extension of Higher Criticism Thresholding based on innovations is proposed
Salloum, Zahraa. "Maximum de vraisemblance empirique pour la détection de changements dans un modèle avec un nombre faible ou très grand de variables." Thesis, Lyon, 2016. http://www.theses.fr/2016LYSE1008/document.
Full textIn this PHD thesis, we propose a nonparametric method based on the empirical likelihood for detecting the change in the parameters of nonlinear regression models and the change in the coefficient of linear regression models, when the number of model variables may increase as the sample size increases. Firstly, we test the null hypothesis of no-change against the alternative of one change in the regression parameters. Under null hypothesis, the consistency and the convergence rate of the regression parameter estimators are proved. The asymptotic distribution of the test statistic under the null hypothesis is obtained, which allows to find the asymptotic critical value. On the other hand, we prove that the proposed test statistic has the asymptotic power equal to 1. The epidemic model, a particular case of model with two change-points, under the alternative hypothesis, is also studied. Afterwards, we use the empirical likelihood method for constructing the confidence regions for the difference between the parameters of a two-phases nonlinear model with random design. We show that the empirical likelihood ratio has an asymptotic χ2 distribu- tion. Empirical likelihood method is also used to construct the confidence regions for the difference between the parameters of a two-phases nonlinear model with response variables missing at randoms (MAR). In order to construct the confidence regions of the parameter in question, we propose three empirical likelihood statistics : empirical likelihood based on complete-case data, weighted empirical likelihood and empirical likelihood with imputed va- lues. We prove that all three empirical likelihood ratios have asymptotically χ2 distributions. An another aim for this thesis is to test the change in the coefficient of linear regres- sion models for high-dimensional model. This amounts to testing the null hypothesis of no change against the alternative of one change in the regression coefficients. Based on the theoretical asymptotic behaviour of the empirical likelihood ratio statistic, we propose, for a deterministic design, a simpler test statistic, easier to use in practice. The asymptotic normality of the proposed test statistic under the null hypothesis is proved, a result which is different from the χ2 law for a model with a fixed variable number. Under alternative hypothesis, the test statistic diverges
Manrique, Tito. "Functional linear regression models : application to high-throughput plant phenotyping functional data." Thesis, Montpellier, 2016. http://www.theses.fr/2016MONTT264/document.
Full textFunctional data analysis (FDA) is a statistical branch that is increasingly being used in many applied scientific fields such as biological experimentation, finance, physics, etc. A reason for this is the use of new data collection technologies that increase the number of observations during a time interval.Functional datasets are realization samples of some random functions which are measurable functions defined on some probability space with values in an infinite dimensional functional space.There are many questions that FDA studies, among which functional linear regression is one of the most studied, both in applications and in methodological development.The objective of this thesis is the study of functional linear regression models when both the covariate X and the response Y are random functions and both of them are time-dependent. In particular we want to address the question of how the history of a random function X influences the current value of another random function Y at any given time t.In order to do this we are mainly interested in three models: the functional concurrent model (FCCM), the functional convolution model (FCVM) and the historical functional linear model. In particular for the FCVM and FCCM we have proposed estimators which are consistent, robust and which are faster to compute compared to others already proposed in the literature.Our estimation method in the FCCM extends the Ridge Regression method developed in the classical linear case to the functional data framework. We prove the probability convergence of this estimator, obtain a rate of convergence and develop an optimal selection procedure of theregularization parameter.The FCVM allows to study the influence of the history of X on Y in a simple way through the convolution. In this case we use the continuous Fourier transform operator to define an estimator of the functional coefficient. This operator transforms the convolution model into a FCCM associated in the frequency domain. The consistency and rate of convergence of the estimator are derived from the FCCM.The FCVM can be generalized to the historical functional linear model, which is itself a particular case of the fully functional linear model. Thanks to this we have used the Karhunen–Loève estimator of the historical kernel. The related question about the estimation of the covariance operator of the noise in the fully functional linear model is also treated.Finally we use all the aforementioned models to study the interaction between Vapour Pressure Deficit (VPD) and Leaf Elongation Rate (LER) curves. This kind of data is obtained with high-throughput plant phenotyping platform and is well suited to be studied with FDA methods
Coudin, Élise. "Inférence exacte et non paramétrique dans les modèles de régression et les modèles structurels en présence d'hétéroscédasticité de forme arbitraire." Thèse, Paris, EHESS, 2007. http://hdl.handle.net/1866/1506.
Full textLabit, Yann. "Contribution à la commande non linéaire par des approches linéaires." Phd thesis, INSA de Toulouse, 2002. http://tel.archives-ouvertes.fr/tel-00131792.
Full textAhmed, Mohamed Salem. "Contribution à la statistique spatiale et l'analyse de données fonctionnelles." Thesis, Lille 3, 2017. http://www.theses.fr/2017LIL30047/document.
Full textThis thesis is about statistical inference for spatial and/or functional data. Indeed, weare interested in estimation of unknown parameters of some models from random or nonrandom(stratified) samples composed of independent or spatially dependent variables.The specificity of the proposed methods lies in the fact that they take into considerationthe considered sample nature (stratified or spatial sample).We begin by studying data valued in a space of infinite dimension or so-called ”functionaldata”. First, we study a functional binary choice model explored in a case-controlor choice-based sample design context. The specificity of this study is that the proposedmethod takes into account the sampling scheme. We describe a conditional likelihoodfunction under the sampling distribution and a reduction of dimension strategy to definea feasible conditional maximum likelihood estimator of the model. Asymptotic propertiesof the proposed estimates as well as their application to simulated and real data are given.Secondly, we explore a functional linear autoregressive spatial model whose particularityis on the functional nature of the explanatory variable and the structure of the spatialdependence. The estimation procedure consists of reducing the infinite dimension of thefunctional variable and maximizing a quasi-likelihood function. We establish the consistencyand asymptotic normality of the estimator. The usefulness of the methodology isillustrated via simulations and an application to some real data.In the second part of the thesis, we address some estimation and prediction problemsof real random spatial variables. We start by generalizing the k-nearest neighbors method,namely k-NN, to predict a spatial process at non-observed locations using some covariates.The specificity of the proposed k-NN predictor lies in the fact that it is flexible and allowsa number of heterogeneity in the covariate. We establish the almost complete convergencewith rates of the spatial predictor whose performance is ensured by an application oversimulated and environmental data. In addition, we generalize the partially linear probitmodel of independent data to the spatial case. We use a linear process for disturbancesallowing various spatial dependencies and propose a semiparametric estimation approachbased on weighted likelihood and generalized method of moments methods. We establishthe consistency and asymptotic distribution of the proposed estimators and investigate thefinite sample performance of the estimators on simulated data. We end by an applicationof spatial binary choice models to identify UADT (Upper aerodigestive tract) cancer riskfactors in the north region of France which displays the highest rates of such cancerincidence and mortality of the country
Hamadeh, Tawfik. "Inférence statistique de modèles GARCH non linéaires." Lille 3, 2010. http://www.theses.fr/2010LIL30048.
Full textThis thesis is devoted to the statistical inference of two wide classes of non linear GARCH models. Firstly, several estimation methods of a class of power-transformed treshold GARCH models are considered in two situations. When the power of the transformation is known, the asymptotic properties of the quasi-maximum likelihood estimator (QMLE) are established under mild conditions. Two sequences of least-squares estimators are also considered in the pure ARCH case, and it is shown that they can be asymptotically more accurate than the QMLE for certain power transformations. In the case where the power of the transformation is jointly estimated with others parameters, the asymptotic properties of the QMLE are proven under the assumption that the noise has a density. Moreover, we establish the consistency and the asymptotic normality of a class of non-gaussian QML estimators in the case where alternatives to the classical QML estimator, especially, when the rescaled errors are heavy tailed. In the second part of this thesis, we introduce a general class of weak GARCH processes with contains a large family of volability models. This representation consists of two ARMA equations, the first one on the observed process and the second one on a function of its linear innovation. Under some moment conditions, strong mixing and stationarity assumptions, the asymptotic properties of two-stage least-squares estimator for the proposed model are established. We also consider the estimation of the asymptotic covariance matrix of this estimator
Buatois, Simon. "Novel pharmacometric methods to improve clinical drug development in progressive diseases." Thesis, Sorbonne Paris Cité, 2018. http://www.theses.fr/2018USPCC133.
Full textIn the mid-1990, model-based approaches were mainly used as supporting tools for drug development. Restricted to the “rescue mode” in situations of drug development failure, the impact of model-based approaches was relatively limited. Nowadays, the merits of these approaches are widely recognised by stakeholders in healthcare and have a crucial role in drug development for progressive diseases. Despite their numerous advantages, model-based approaches present important drawbacks limiting their use in confirmatory trials. Traditional pharmacometric (PMX) analyses relies on model selection, and consequently ignores model structure uncertainty when generating statistical inference. The problem of model selection is potentially leading to over-optimistic confidence intervals and resulting in a type I error inflation. Two projects of this thesis aimed at investigating the value of innovative PMX approaches to address part of these shortcomings in a hypothetical dose-finding study for a progressive disorder. The model averaging approach coupled to a combined likelihood ratio test showed promising results and represents an additional step towards the use of PMX for primary analysis in dose-finding studies. In the learning phase, PMX is a key discipline with applications at every stage of drug development to gain insight into drug, mechanism and disease characteristics with the ultimate goal to aid efficient drug development. In this thesis, the merits of PMX analysis were evaluated, in the context of Parkinson’s disease. An item-response theory longitudinal model was successfully developed to precisely describe the disease progression of Parkinson’s disease patients while acknowledging the composite nature of a patient-reported outcome. To conclude, this thesis enhances the use of PMX to aid efficient drug development and/or regulatory decisions in drug development
Paccard, Caroline. "Développement d'outils statistiques pour la mise en place de boucles de régulation en microélectronique." Toulouse 3, 2008. http://thesesups.ups-tlse.fr/536/.
Full textIn semiconductor manufacturing, classical process control is no sufficient anymore for new technologies. A more accurate control can be achieved with closed-loop control (run-to-run). This thesis designs a statistical methodology aimed at deploying closed-loop control in semiconductor manufacturing. This methodology remains general and can be easily transposed to other industries. Studying closed-loop control, we have come to the issue of measurement reliability. Thus we have created a new indicator of measurement variability, called global capability, which can be applied when one parameter is measured by several metrology tools. An operational solution has been proposed through a software creation. It has been implemented and put into production to compute this new indicator. After its definition, the methodology for closed-loop control design has been applied to a polishing process. It has conducted us to an original process modeling thanks to a linear mixed model. We have also compared and optimized several regulation algorithms (EWMA, double EWMA, Kalman filter. . . ). For cost reasons, the considered regulation algorithms could not be all tested and compared in production. As a result, we have designed a process simulation based on production data and on a process modeling. This simulation can predict and compare what will be the regulation algorithm behavior in production. For the polishing process, an optimal algorithm has been chosen
Semenou, Michel. "Construction de plans expérimentaux et propriétés de modèles linéaires généralisés mal spécifiés : application à une étude de fiabilité." Toulouse 3, 1994. http://www.theses.fr/1994TOU30006.
Full textHomkham, Nontiya. "Long-Term Evolution Of Lipids In Thai HIV-Infected Patients On Treatment." Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLS094.
Full textAs other antiretroviral drugs, treatment with efavirenz has been associated with potentially unfavorable lipid profile changes in adults and in children. The thesis addressed the question of whether these changes depend on efavirenz plasma concentrations and if dose adjustments could be envisioned without loss of efficacy.To estimate individual efavirenz exposure over 24 hours, a population pharmacokinetic model was developed using data from HIV infected children. Simulations for a normalized population receiving efavirenz dosed according recommendations predicted that 15% of children would have insufficient mid dose concentrations, associated with a 23% risk of viral replication.To describe the relationship between efavirenz concentrations and cholesterol changes, population pharmacokinetic-pharmacodynamic (PK-PD) indirect response models were developed. The selected model predicted that individual efavirenz concentrations were associated with an increase in high-density lipoprotein concentrations over 5 years but with an increase in low-density lipoprotein concentrations only during the first 4 months of treatment followed by a gradual return to baseline.To study the importance of efavirenz concentrations with regard to efficacy, a PK-PD dynamics model was developed to describe the relationship between concentrations and HIV RNA load and CD4 cell count evolutions. A score was defined based on a pharmacodynamic hypothesis to predict the risk of viral replication.Using US Food and Drug Administration dosing recommendations in children ensure optimal efficacy and potentially favorable changes in cholesterol fractions
Valade, Charles. "Développement d'une méthodologie adaptée à l'industrie microélectronique pour la reconstruction topographique par imagerie SEM à faisceau inclinable." Thesis, Université Grenoble Alpes, 2020. http://www.theses.fr/2020GRALT015.
Full textWith the advancement of microelectronics technologies, the architecture of electronic components is becoming increasingly complicated. However, knowledge of the dimensional characteristics of the structures is important in order to be able to understand and optimize the behavior of these components. This is why there is a need to develop rapid, non-destructive three-dimensional measurement methods.The scanning electron microscope (SEM) is widely used to carry out dimensional measurements because it responds to the problems of speed and non-destructivity. However, obtaining quantitative and precise three-dimensional information is a challenge.Thanks to an electron microscope whose electron beam can be tilted, it is possible to obtain images at different viewing angles. From the analysis of these images, the height and the sidewall angles of the observed pattern can be determined geometrically.However, since electronic imaging is the result of electron-matter interactions, it is important to understand the origin of the formation of SEM images, in order to be able to analyze them correctly. This is why a study was carried out using physical simulation software to observe and understand the impact of the topography of a pattern on the resulting SEM image.From these observations, metrics were created on the SEM images to analyze them quantitatively.A linear model was then created using physical simulations to estimate the topographic quantities from these metrics. It was then calibrated on real SEM measurements, by comparing them to three-dimensional reference measurements by atomic force microscopy (AFM). This model was created for the reconstruction of “line” type patterns in etched silicon. Thanks to this model, reconstructions of real patterns were made. Finally, work was started on the creation of a model for "trench" and "dense" type patterns in etched silicon
Alata, Olivier. "Contributions à la description de signaux, d'images et de volumes par l'approche probabiliste et statistique." Habilitation à diriger des recherches, Université de Poitiers, 2010. http://tel.archives-ouvertes.fr/tel-00573224.
Full textPadilla, Arturo. "Identification récursive de systèmes continus à paramètres variables dans le temps." Thesis, Université de Lorraine, 2017. http://www.theses.fr/2017LORR0119/document.
Full textThe work presented in this thesis deals with the identification of dynamic systems represented through continuous-time linear models with slowly time-varying parameters. The complexity of the identification problem comes on the one hand from the unknown character of the parameter variations and on the other hand from the presence of noises of unknown nature on the measured signals. The proposed solutions rely on a judicious combination of the Kalman filter assuming that the variations of the parameters can be represented in the form of a random walk, and the method of the instrumental variable which has the advantage of being robust with respect to the nature of the measurement noises. The recursive algorithms are developed in an open-loop and closed-loop identification setting. The different variants are distinguished by the way in which the instrumental variable is built. Inspired by the solution developed for time-invariant linear systems, an adaptive construction of the instrumental variable is suggested in order to be able to follow the evolution of the parameters as well as possible. The performance of the developed methods are evaluated using Monte Carlo simulations and show the supremacy of the proposed solutions based on the instrumental variable compared with the more classical least squares based approaches. The practical aspects and implementation issues are of paramount importance to obtain a good performance when these estimators are used. These aspects are studied in detail and several solutions are proposed not only to robustify the estimators with respect to the choice of hyperparameters but also with respect to their numerical implementation. The algorithms developed have enhanced the functions of the CONTSID toolbox for Matlab. Finally, the developed estimators are considered in order to track parameters of two physical systems: a benchmark available in the literature consisting of a bandpass electronic filter and a throttle valve equipping the car engines. Both applications show the potential of the proposed approaches to track physical parameters that vary slowly over time
Hamouda, Makram. "Perturbations singulières pour des EDP linéaires et non linéaires en présence de singularités." Paris 11, 2001. https://tel.archives-ouvertes.fr/tel-00001931.
Full textIn my thesis, we study some singular perturbation problems (i. E characterized by the presence of a small parameter) which develop boundary layers in some conditions more delicate than usually, namely when the limit solution is not regular. I consider then two classes of regular problems associated to Laplacian and bilaplacian, and a nonlinear problem derived from the Plateau problem (minimal surfaces), for which the limit function has an infinite normal derivative on some parts of the boundary of the domain. The first part of this thesis is concerned with the study of two singular linear models associated to non classical singular perturbations. In fact, the presence of a small parameter in the partial differential equations involve the appearance of classical boundary layers near the boundary for the regularized solution. However, if we consider moreover a discontinuous source function (even a distribution), we note the appearance of non classical boundary layers in the interior of the domain; the study of these boundary layers consists the main object of this first part. In the second part of my thesis, i am interesting to study the minimal surfaces problem on a domain formed by two concentric circles. For some boundary data, this problem do not have any solution and its weak solution called "the generalized solution" has an infinite gradient. To solve this difficulty, we introduce an elliptic regularsation which involve boundary layers near the boundary. The main result of this part consists to give some representation formulas of the solutions, of some approximate solutions, and some estimates of the order of convergence
Vera, Carine. "Modèles linéaires mixtes multiphasiques pour l'analyse de données longitudinales : Application à la croissance des plantes." Montpellier 2, 2004. http://www.theses.fr/2004MON20161.
Full textMarin, Jean-Michel. "Statistiques des modèles à structure de covariance bande-diagonale linéaire." Toulouse 3, 2001. http://www.theses.fr/2001TOU30123.
Full textLe, Thi Xuan Mai. "Estimation semi-paramétrique et application à l’évaluation de la biomasse d'anchois." Thesis, Toulouse, INSA, 2010. http://www.theses.fr/2010ISAT0006/document.
Full textThe motivation of this study is to evaluate the anchovy biomass, that is estimate the egg densities at the spawning time and the mortality rate. The data are the anchovy egg densities that are the egg weights by area unit, collected in the Gascogne bay. The problem we are faced is to estimate from these data the egg densities at the spawning time. Until now, this is done by using the classical exponential mortality model. However, such model is inadequate for the data under consideration because of the great spatial variability of the egg densities at the spawning time. They are samples of generated by a r.v whose mathematical expectation is a0 and the probability density function is fA. Therefore, we propose an extended exponential mortality model Y (tj,kj) = A (tj,kj) e-z0tj +e(tj,kj) where A(tj,kj) and e(tj,kj) are i.i.d, with the random variables A and e being assumed to be independent. Then the problem consists in estimating the mortality rate and the probability density of the random variable . We solve this semiparametric estimation problem in two steps. First, we estimate the mortality rate by fitting an exponential mortality model to averaged data. Second, we estimate the density fA by combining nonparametric estimation method with deconvolution technique and estimate the parameter z0. Theoretical results of consistence of these estimates are corroborated by simulation studies
Launay, Tristan. "Méthodes Bayésiennes pour la prévision de consommation d’électricité." Nantes, 2012. http://www.theses.fr/2012NANT2074.
Full textIn this manuscript, we develop Bayesian statistics tools to forecast the French electricity load. We first prove the asymptotic normality of the posterior distribution (Bernstein-von Mises theorem) for the piecewise linear regression model used to describe the heating effect and the consistency of the Bayes estimator. We then build a a hierarchical informative prior to help improve the quality of the predictions for a high dimension model with a short dataset. We typically show, with two examples involving the non metered EDF customers, that the method we propose allows a more robust estimation of the model with regard to the lack of data. Finally, we study a new nonlinear dynamic model to predict the electricity load online. We develop a particle filter algorithm to estimate the model et compare the predictions obtained with operationnal predictions from EDF
Saidi, Youssef. "Etude probabiliste et statistique de modèles conditionnellement hétéroscédastiques non linéaires." Lille 3, 2003. http://www.theses.fr/2003LIL30020.
Full textVauglin, François. "Modèles statistiques des imprécisions géométriques des objets géographiques linéaires." Université de Marne-la-Vallée, 1997. http://www.theses.fr/1997MARN0010.
Full textKhaled, Mohamad. "Estimation bayésienne de modèles espace-état non linéaires." Paris 1, 2008. http://www.theses.fr/2008PA010048.
Full text