Dissertations / Theses on the topic 'Méthode longitudinale'
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 'Méthode longitudinale.'
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.
Arnal, Mathieu. "Développement d'une évaluation génomique pour l'analyse de données longitudinales : application aux contrôles élémentaires chez les caprins laitiers." Thesis, Toulouse, INPT, 2019. http://www.theses.fr/2019INPT0124.
Full textGenetic improvement of dairy goats is based on the measurement of the quantity and quality of milk production of females on farms, at intervals of 4 to 5 weeks during lactation, according to strict protocols. The genetic evaluation of the quantity of milk is based on the estimation of the total quantity of milk produced per lactation. This selection based on the total quantity of milk in lactation tends to select animals with increasingly high peak lactation production. High production at the beginning of lactation can cause metabolic problems for females. In addition, in the case of dairy goats, in a context of seasonal production, a milk production that is maintained after the peak, i.e. persistent, would allow a more spread out milk production, in line with market expectations. There is therefore a zootechnical and economic interest in wanting to select more persistent dairy goats. In our study, the approach consists in modelling the shape of the lactation curve based on the information collected during each farm test. Models allowing the analysis of such longitudinal data are generally called test-day models. One of the main interests is to take better account of environmental effects, affecting production on test-day, with a herd-test-day effect depending only on the animals present during the test. The second advantage of this type of model is that most genetic and environmental effects are modelled as curves, so it would be possible to select animals with the best genetic value for persistence. The development of these models requires the prior study of the environmental effects affecting milk production over time. Following a detailed descriptive analysis of the lactation curves of the two main French goat breeds (Alpine and Saanen), we showed that there was a variability in the shape of the lactation curves, and in particular the month of calving was involved in the different curve shapes. Then we proposed a random regression model, similar to that developed in French dairy cattle. The proposed modeling makes it possible to obtain two genetic indexes directly: one corresponding to the genetic value of the animal for the total quantity of milk during lactation and a second corresponding to a genetic value of the animal's milk persistency, without correlation between the two. The model proposed is more relevant than the current one because it takes into account, in a disjointed way, the goats in primiparous and the goats in multiparous. We also studied correlations between indexes of different traits during lactation and correlations between persistence and AI fertility or between persistence and longevity. In the last part of the thesis, we extended the genetic evaluation of test-day to a genomic evaluation model (Single-step GBLUP) allowing to exploit all available molecular information (genotyping SNP 50K). A validation of this model and a comparison with the current model was carried out. The main difference between the Single-step GBLUP RRM and the Single-step GBLUP lactation model currently in use was the differences in the averages of the estimated indexes per year of birth of the bucks. Finally, from the Single-step GBLUP model we have identified some interested regions of the genome linked to milk persistence
Turc, Ioan Emil. "Le changement radical et le temps : étude sur l'accélération des transitions organisationnelles." Aix-Marseille 3, 2003. http://www.theses.fr/2003AIX32039.
Full textThe excessive duration of radical organizational change becomes more and more unacceptable to executives. However, organizational change research is more preoccupied with refining the implementation strategies for radical change, rather than seeking solutions to accelerate the process. This issue is addresses by the present research, whose aim is to answer the following question : what factors and processes accelerate radical organizational change ? Based on an exploratory research design of comparative, longitudinal and retrospective studies, this thesis unveils 10 main principles of acceleration. The explanatory theories developed there of highlight the roles of crisis perceptions and trust, commitment and conformity, activity planning and political management in the elaboration of effective acceleration techniques
Pennec, Sophie. "Applications démo-économiques de la méthode des microsimulations." Paris, Institut d'études politiques, 1994. http://www.theses.fr/1994IEPP0031.
Full textThis work deals with several demo-economic applications of the method of microsimulations. This method have been first used at the end of the 50's. The first chapter, first of all, describes the main principles of microsimulation, then set out the different types of models based on this method and explain the main advantages offered by it. Then, some existing models are presented in a more detailed way so as to show the logic behing their design. Three applications will be the object of the following chapters. The first one is purely demographic. It covered the development of the families with four generations. Taking the longitudinal approach which allow the use of microsimulation, it aims to measure the influence of the recent and future changes of fertility and mortality, at the level of its intensity as well as at that of its timing, and the evolution of the probability to belong in this sort of family. The second application resumes the longitudinal perspective to describe a double biography of women during their retirement; their family life (presence or not of the spouse, of the children) and the changes in income during this period. The first biography gives a measure of the isolation phenomena as they got older, isolation linked mainly to mortality. The second biography, made possible thanks to microsimulation, gives the financial situation of the persons when they retire. The demography of a firm is the object of the last application. It shows, first of all, how microsimulation can be applied to different human resources management methods and then which are the implications of the use of these defined ways of management on the evolution of the firm
Feydy, Antoine. "Plasticité cérébrale et récupération motrice après un accident vasculaire cérébral ischémique : étude en imagerie par résonance magnétique fonctionnelle (IRMf)." Paris 13, 2002. http://www.theses.fr/2002PA132018.
Full textBodein, Antoine. "Mise en place d'approches bioinformatiques innovantes pour l'intégration de données multi-omiques longitudinales." Doctoral thesis, Université Laval, 2021. http://hdl.handle.net/20.500.11794/69592.
Full textNew high-throughput «omics» technologies, including genomics, epigenomics, transcriptomics, proteomics, metabolomics and metagenomics, have expanded considerably in recent years. Independently, each omics technology is an essential source of knowledge for the study of the human genome, epigenome, transcriptome, proteome, metabolome, and also its microbiota, thus making it possible to identify biomarkers leading to diseases, to identify therapeutic targets, to establish preventive diagnoses and to increase knowledge of living organisms. Cost reduction and ease of multi-omics data acquisition resulted in new experimental designs based on time series in which the same biological sample is sequenced, measured and quantified at several measurement times. Thanks to the combined study of omics technologies and time series, it is possible to capture the changes in expression that take place in a dynamic system for each molecule and get a comprehensive view of the multi-omics interactions, which was inaccessible with a simple standard omics approach. However, dealing with this amount of multi-omics data faces new challenges: continuous technological evolution, large volumes of produced data, heterogeneity, variety of omics data and interpretation of integration results require new analysis methods and innovative tools, capable of identifying useful elements through this multitude of information. In this perspective, we propose several tools and methods to face the challenges related to the integration and interpretation of these particular multi-omics data. Finally, integration of longidinal multi-omics data offers prospects in fields such as precision medicine or for environmental and industrial applications. Democratisation of multi-omics analyses and the implementation of innovative integration and interpretation methods will definitely lead to a deeper understanding of eco-systems biology.
Tjani, Essama Ndjeng Jeanne Chantal. "Étude longitudinale du statut nutritionnel et vitaminique de sujets âgés récemment admis en institution." Dijon, 2000. http://www.theses.fr/2000DIJOMU07.
Full textRomeo, Silvia. "La construction de la représentation d'action complexe au cours du développement par la médiation des langues italienne et française : étude longitudinale." Paris 10, 2001. http://www.theses.fr/2001PA100073.
Full textThis longitudinal research, based on the natural discourse of an Italian French bilingual child in family interactions, follows a developmental, interactional perspective and a conceptual approach. Analysis is based on linguistic data taken from daily situations. Cognitif and linguistic development are considered in terms of the contribution of the coexisting Italian and French linguistic systems. We have considered language acquisition by the mediation of two first languages for the construction of the representation of complex actions, taken as integrating the notional fields of temporality , causal and goal relations. Two developmental stages are considered : from 2 to 3;3 and from 3;4 to 4;9. At the first stage we see the gradual construction of the semantic roles of agent, patient, and instrument and of the categories of state, action and event as related to temporal, causal and goal relations. These relations develop in the second stage "transformational system" organisations. First stage results confirm the general hypothesis that the acquisition of two first languages is a semantic and formal factor of organisation, as based on a more complex anchoring points than first language acquisition. Second stage results show that child productions, as linearization of hierarchical series of actions of a high degree of temporal granularity, reveal an on line planification, as founded on a scheme of complex action representation. We have taken into consideration the role of two first languages in the organisation of knowledge and the way of structuring it
Minini, Pascal. "Modélisation des observations longitudinales incomplètes." Paris 11, 2004. http://www.theses.fr/2004PA11T060.
Full textIn longitudinal studies, subjects are repeatedly observed to obtain measurements of some response. The protocol of such studies plans to collect a fixed number of responses for each subject, during a predefined follow-up period. However, it is extremely rare that all the planned measurements are actually performed. The analysis of incomplete data has become a major topic in statistics during the last years. Many methods have been proposed, but it is generally impossible to check their validity. It is now recommended to perform a sensitivity analysis, to evaluate the extent to which the results from a study can be affected by different assumptions regarding the missing data process. This report highlights the need for a sensitivity analysis, and shows how this aim can be achieved in three different situations, normal data, binary data and survival data
Charton, Emilie. "Analyse longitudinale des données de qualité de vie relative à la santé en cancérologie : vers une standardisation de la méthode du temps jusqu’à détérioration." Thesis, Bourgogne Franche-Comté, 2020. http://www.theses.fr/2020UBFCE003.
Full textThe purpose/aim of this thesis is to contribute to the analysis and comparison of PROs (« patient-reported outcomes ») data in oncology clinical trials. The interpretation of such results remains complex and unstandardized. One of the many ways to carry out a longitudinal analysis of PRO data is the time to deterioration (TTD) approach. Within of the scope of this project, some of the research examined which definitions of TTD are used and pointed out that some recommendations have not been followed. Moreover, due to the variability of the definitions in use, the comparison of various results from clinical trials is compromised. A clear definition of what is considered to be a « deterioration » is required for the TTD approach. It will depend on many criteria such as the location of the cancer, the therapeutic setting, the reference score, the minimal important difference perceived by the patient, as well as on censoring rules. Two SAS macros were developed on the TTD method as a way to optimize and harmonize the TTD definitions that are being used, as well as to be able to have comparable results and consequently a way to help standardize those definitions. In this perspective, a study conducted on a cohort of adjuvant breast cancer patients led to more focus on the first deterioration of the patient and the management of non-randomization at baseline. In parallel, this method was also implemented for a randomized phase II trial on patients with metastatic pancreatic cancer. During this trial, the impact of the occurrence of missing data at baseline was handled by applying a multiple imputation based on the Markov Chain Monte Carlo method. These works highlight the need to continue developing a consensus for the longitudinal analysis of PROs data in oncology clinical trials
Trarieux, Chloé. "Mesure des propriétés viscoélastiques non linéaires par une méthode d'acousto-élasticité dynamique : application aux produits cosmétiques." Thesis, Tours, 2014. http://www.theses.fr/2014TOUR3312/document.
Full textFew tools have been developed for industrial quality control of textures on production lines. The use of contactless techniques. based on acoustic wavcs, offers an obvious advantage in food-processing industry and cosmetics. The dynamic acoustoelastic testing (DAET) is based on the interaction between a low-frequency compressionlexpansion pump wave and an ultrasound probe wave. leading to the quantification of the viscoelastic properties of the matter. We have initially developed a model describing the variations of the viscoelastic modulus: quantification of nonlinear elastic (B/A, C/A...) and viscous (ωηB/A. ωηC/A...) parameters. The DAET method and related model were then validated in homogeneous media (water, oils and gels) leading to low values of viscoelastic nonlineaiities (B/A<15 and ωηB/A<1), essentially governed by the fluid nature or state change. However, the most significant results were obtained in granular or air-based media (dry powder, foam and hollow beads): high values of quadratic and cubic nonlinearities due to microinhomogeneitics. This method appears to be an interesting alternative to conventional rheometry, especially for the characterization of these complex fluids
Baghdadli, Amaria. "Étude des facteurs de variabilité des troubles autistiques de l'enfant : vers une identification de facteurs pronostiques de l'autisme." Montpellier 1, 2001. http://www.theses.fr/2001MON1T009.
Full textThelisson, Anne-Sophie. "Intégration post-fusion : une lecture paradoxale comme moyen pour comprendre le processus d'intégration : éclairages par une étude de cas longitudinale in vivo." Thesis, Aix-Marseille, 2017. http://www.theses.fr/2017AIXM0135.
Full textDespite the number of mergers and acquisitions (M&As), their high failure rate calls to continue research by proposing new reading keys. Many studies argue that M&As require new approaches to understand the organizational complexity and dynamism of these operations, and more specifically concerning post-merger integration (PMI) process. This phase is defined as a crucial one because it determines the success or failure of the merger. Our research provides an analysis of the dynamics inherent in the process, and especially during the PMI. This study is based on an integrative logic, considering the multiplicity of dynamics at work to understand how they engender or hinder the success of PMI. We use these dynamics as a means to capture the complexity inherent in the merger process, and as an opportunity to capture the dynamics of PMI. The concept of paradox allows us to make intelligible these inter-related dynamics within a defined temporal framework. Paradoxes provides a framework to decipher the dynamics inherent in organizations. The paradoxical reading allows a comprehensive analysis of the dynamics at work in the merger process. This leads us to ask the following research question: "How does the management of paradoxes benefits post-merger integration? ". The thesis is based on a qualitative methodology of a longitudinal and real-time case study of a 24-month merger, once the operation has been signed. It emerges from the thesis that the evolutionary nature of paradoxes during the post-merger integration, as well as their multiplicity (categories, actors, levels), allows us to understand how these dynamics interfere in the success of the PMI process
Sort, Lucas. "Development of new statistical approaches for the investigation of multiblock and tensor longitudinal data." Electronic Thesis or Diss., université Paris-Saclay, 2025. http://www.theses.fr/2025UPASG001.
Full textMedical researchers have been confronted with increasingly complex data over the past decades. Indeed, data is now often observed on different occasions over time to help track the evolution of biological processes that can provide medical information. In this context, this data, which can be referred to as longitudinal data, often comes with a complex and high-dimensional structure. For example, in an increasing number of cases, it may be structured as a multidimensional array, also known as a tensor. Similarly, it may be organized into several blocks, especially in the multimodal setting. The statistical analysis of such structured data poses several methodological challenges. Integrating the tensor or multiblock structure or considering the properties associated with longitudinal sampling seems essential to provide a relevant and more accurate study of the data. In this context, we suggest using the functional data analysis framework to adapt and extend several multivariate statistical approaches to the longitudinal setting. First, we propose to extend the multiblock analysis framework of Regularized Generalized Canonical Correlation Analysis to allow researchers to explore associations between multiple longitudinal markers. Then, in the tensor setting, we introduce a new tensor decomposition model to extract the most relevant information from high-dimensional longitudinally sampled tensors. Finally, we propose a new model for regressing a scalar response from longitudinally sampled tensor-structured covariates. Several applications are considered to illustrate the numerous settings in which these methodological developments could be used to assist medical researchers
Dantzer, Cécile Elsa. "Anxiété, dépression et facteurs psychologiques chez les adolescents malades chroniques." Bordeaux 2, 2002. http://www.theses.fr/2002BOR20974.
Full textLange, Marie. "Impact de la chimiothérapie sur les fonctions cognitives de patientes âgées traitées pour un cancer du sein localisé." Caen, 2014. http://www.theses.fr/2014CAEN1017.
Full textWhile frequent cognitive complaints after chemotherapy for breast cancer, very few studies have investigated this issue among old patients, though they are at greater risk to have cancer and to develop cognitive impairment. The main aim of this thesis was to better understand the effect of adjuvant chemotherapy, and, to a larger extent, the effect of cancer on cognition in early-stage breast cancer patients over 65. For this purpose, a longitudinal study with pre- and post-adjuvant treatment assessments was realized to evaluate the incidence of cognitive impairment and to study the associated variables. We showed that 41% of patients presented objective cognitive impairment before any adjuvant treatment. This impairment was neither related to mood, fatigue, cancer stage, nor to geriatric assessment. After adjuvant treatment and regardless of its type, 49% of patients had cognitive decline, which mainly regarded working memory. Among patients who had cognitive decline, 64% had pathological decline. Adjuvant chemotherapy gave rise to a significant increase of cognitive complaints. Moreover, the oldest patients treated with chemotherapy tended to be at higher risk for developing cognitive decline, especially when they had received docetaxel. Age and chemotherapy are therefore two important factors to be considered in the development of cognitive impairment in early-stage breast cancer patients
Boukadida, Nahed. "Connaissances phonologiques et morphologiques dérivationnelles et apprentissage de la lecture en arabe (Etude longitudinale)." Phd thesis, Université Rennes 2, 2008. http://tel.archives-ouvertes.fr/tel-00300544.
Full textEssig, Elena. "Les facteurs de développement de la propriété psychologique : une étude longitudinale à l'armée de l'air." Thesis, Tours, 2018. http://www.theses.fr/2018TOUR1001/document.
Full textThis work investigates the development of three types of PO - Organizational Psychological Ownership (OPO), the Collective Organizational Psychological Ownership (PPOC) and Psychological Ownership towards the group (POGR) and their fluctuation over time in a public organization - the French Air Force. We test a model using structural equation modeling. The fluctuations in PO levels are tested using analysis of variances. The quantitative data were collected from three cohorts of 100 non-commissioned officers each, at three different phases of the military training process. The results of the study show a variability of the antecedents of the three types of PO that depend on the training phases. Even though the OPO and COPO have similar antecedents, they do not evolve in the same way. The feelings of POGR are most developed and depend on the group cohesion. There are significant downward fluctuations in PO levels over time
Wang, Dawei. "Mesoscopic modeling and simulation on the forming process of textile composites." Thesis, Lyon, 2016. http://www.theses.fr/2016LYSEI108/document.
Full textThis thesis is devoted to the mesoscopic study on the performance of textile reinforcements. F.E. simulation is carried out on a mesoscale model for the fibrous material, based on which two kinds of new deformation modes are developed. The first one is a longitudinal compression mode, which is used to reflect the small stiffness when the yarn is compressed longitudinally. The incompatibility problem between the small longitudinal compression stiffness and the large tension stiffness are solved by three different strategies: constraining the critical step time, adding the nonlinear tension part, or using a new strategy to update the stress. The second one is transverse expansion mode that could reflect the influence from longitudinal deformation to transverse deformation. This deformation could be found in tomography view but was ignored by the former researches. An experiment is designed to measure the expansion magnitude, and the geometrical inverse fitting process is applied to measure the value of the longitudinal-transverse Poisson ratio. The parameters of the mesoscale model are measured by a series of mechanical experiments and the simulation results are verified by the tomography methodology
Schiratti, Jean-Baptiste. "Methods and algorithms to learn spatio-temporal changes from longitudinal manifold-valued observations." Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLX009/document.
Full textWe propose a generic Bayesian mixed-effects model to estimate the temporal progression of a biological phenomenon from manifold-valued observations obtained at multiple time points for an individual or group of individuals. The progression is modeled by continuous trajectories in the space of measurements, which is assumed to be a Riemannian manifold. The group-average trajectory is defined by the fixed effects of the model. To define the individual trajectories, we introduced the notion of « parallel variations » of a curve on a Riemannian manifold. For each individual, the individual trajectory is constructed by considering a parallel variation of the average trajectory and reparametrizing this parallel in time. The subject specific spatiotemporal transformations, namely parallel variation and time reparametrization, are defined by the individual random effects and allow to quantify the changes in direction and pace at which the trajectories are followed. The framework of Riemannian geometry allows the model to be used with any kind of measurements with smooth constraints. A stochastic version of the Expectation-Maximization algorithm, the Monte Carlo Markov Chains Stochastic Approximation EM algorithm (MCMC-SAEM), is used to produce produce maximum a posteriori estimates of the parameters. The use of the MCMC-SAEM together with a numerical scheme for the approximation of parallel transport is discussed. In addition to this, the method is validated on synthetic data and in high-dimensional settings. We also provide experimental results obtained on health data
Hurault-Delarue, Caroline. "Approche longitudinale et quantitative de l'exposition aux médicaments dans les études de pharmaco-épidémiologie : développement méthodologique et application aux expositions au cours de la grossesse dans la cohorte EFEMERIS." Thesis, Toulouse 3, 2016. http://www.theses.fr/2016TOU30041.
Full textThe intensity and duration of drug exposure may contribute to the occurrence of drug adverse effects. However, these parameters are rarely simultaneously addressed in studies of risks associated to drug exposure, in particular during pregnancy. Administrative databases give the opportunity to apprehend these parameters and to reconstruct the history of patient drug exposure. The aim of this research is to develop a new method of exposure measurement in order to cluster individuals, taking into account both the intensity and the evolution of exposure. The application to pharmaco-epidemiological studies allow a quantitative approach of drug exposure and longitudinal over time defining individual trajectories of exposure. We used an unsupervised clustering method based on an implementation of K-means adapted to longitudinal data analysis to cluster individuals in homogeneous groups according to their trajectories. This "trajectory method" was applied to psychotropic drug exposure during pregnancy, using EFEMERIS database. The first phase of this application led to the identification of clusters with homogeneous profiles. During the second phase, clusters of women exposed to anxiolytic and hypnotic drugs were used as independent variables to study the effects of in utero exposure to these drugs on newborns and children. The study especially indicates a dose-response relationship between the 4 clusters and an increased risk of neonatal pathologies after an exposure to a heavy drug burden. By contrast, results concerning women punctually exposed or exposed to a light drug burden were reassuring. This application to real-clinical-data has validated this method and demonstrates the interest value of considering intensity and evolution of drug exposure over time in pharmaco-epidemiological studies. The proposed method could be adapted to other populations, classes of drugs and other types of exposure. This "trajectorial" approach of exposure opens up new prospects for future epidemiological studies
Meneghel, Danilevicz Ian. "Robust linear mixed models, alternative methods to quantile regression for panel data, and adaptive LASSO quantile regression with fixed effects." Electronic Thesis or Diss., université Paris-Saclay, 2022. http://www.theses.fr/2022UPAST176.
Full textThis thesis consists of three chapters on longitudinal data analysis. Linear mixed models are discussed, both random effects (where individual intercepts are interpreted as random variables) and fixed effects (where individual intercepts are considered unknown constants, i.e., they must be estimated). Furthermore, robust models (resistant to outliers) and efficient models (with low estimator variability) are proposed in the scope of repeated measures. The second part of the thesis is dedicated to quantile regression, which explores the full conditional distribution of an outcome given its predictors. It introduces a more general method for dealing with heteroscedastic variables and longitudinal data. The first chapter is motivated by evaluating the statistical association between air pollution exposure and children and adolescents' lung ability among six months. A robust linear mixed model combined with an equally robust principal component analysis is proposed to deal with multicollinearity between covariates and the impact of extreme observations on the estimates. Huber and Tukey loss functions (M-estimation examples) are considered to obtain more robust estimators than the least squared function usually used to estimate the parameters of linear mixed models. A finite sample size study is carried out in the case where the covariates follow linear time series models with or without additive outliers. The impact of time correlation and outliers on fixed effect parameter estimates in linear mixed models is investigated. In addition, weights are introduced to reduce the estimates' bias even more. The study of the real data revealed that the robust principal component analysis exhibits three principal components explaining more than 90% of the total variability. The second principal component, which corresponds to particles smaller than 10 microns, significantly affects respiratory capacity. In addition, biological indicators such as passive smoking have a negative and significant effect on children's lung ability. The second chapter analyses fixed effect panel data with three different loss functions. To avoid the number of parameters increases with the sample size, we propose to penalize each regression method with the least absolute shrinkage and selection operator (LASSO). The asymptotic properties of two of these new techniques are established. A Monte Carlo study is performed for homoscedastic and heteroscedastic models. Although the model is more challenging to estimate in the heteroscedastic case for most statistical methods, the proposed methods perform well in both scenarios. This confirms that the proposed quantile regression methods are robust to heteroscedasticity. Their performance is tested on economic panel data from the Organisation for Economic Cooperation and Development (OECD). The objective of the third chapter is to simultaneously restrict the number of individual regression constants and explanatory covariates. In addition to the LASSO, an adaptive LASSO is proposed, which enjoys oracle proprieties, i.e., it owns the asymptotic selection of the true model if it exists, and it has the classical asymptotic normality property. Monte Carlo simulations are performed in the case of low dimensionality (much more observations than parameters) and in the case of moderate dimensionality (equivalent number of observations and parameters). In both cases, the adaptive method performs much better than the non-adaptive methods. Finally, we apply our methodology to a cohort dataset of moderate dimensionality. For each chapter, open-source software is written, which is available to the scientific community
Esta tese consiste em três capítulos sobre análise de dados longitudinais. São discutidos modelos lineares mistos, tanto efeitos aleatórios (onde interseptos individuais são interpretados como variáveis aleatórias) quanto efeitos fixos (onde interseptos individuais são considerados constantes desconhecidas, ou seja, devem ser estimadas). Além disso, modelos robustos (resistentes a outliers) e modelos eficientes (com baixa variabilidade de estimadores) são propostos no âmbito de medidas repetidas. A segunda parte da tese é dedicada à regressão quantílica, que explora toda a distribuição condicional de uma variável resposta dado suas preditoras. Ela introduz um método mais geral para lidar com variáveis heterocedásticas e dados longitudinais. O primeiro capítulo é motivado pela avaliação da associação estatística entre a exposição à poluição do ar e a capacidade pulmonar de crianças e adolescentes durante um período de seis meses. Um modelo linear misto robusto combinado com uma análise de componentes principais igualmente robusta é proposto para lidar com a multicolinearidade entre covariáveis e o impacto de observações extremas sobre as estimativas. As funções de perda Huber e Tukey (exemplos de \textit{M-estimation}) são consideradas para obter estimadores mais robustos do que a função de mínimos quadrados geralmente usada para estimar os parâmetros de modelos lineares mistos. Um estudo de tamanho de amostra finito é realizado no caso em que as covariáveis seguem modelos de séries temporais lineares com ou sem outliers aditivos. É investigado o impacto da correlação temporal e outliers nas estimativas de parâmetros de efeito fixo em modelos lineares mistos. Além disso, foram introduzidos pesos para reduzir ainda mais o enviesamento das estimativas. Um estudo em dados reais revelou que a análise robusta dos componentes principais apresenta três componentes principais que explicam mais de 90% da variabilidade total. O segundo componente principal, que corresponde a partículas menores que 10 micrômetros, afeta significativamente a capacidade respiratória. Além disso, os indicadores biológicos como o tabagismo passivo têm um efeito negativo e significativo na capacidade pulmonar das crianças. O segundo capítulo analisa dados de painel com efeito fixo com três diferentes funções de perda. Para evitar que o número de parâmetros aumente com o tamanho da amostra, propomos penalizar cada método de regressão com least absolute shrinkage and selection operator (LASSO). As propriedades assimptóticas de duas dessas novas técnicas são estabelecidas. Um estudo de Monte Carlo é realizado para modelos homocedásticos e heterosecásticos. Embora o modelo seja mais difícil de estimar no caso heterocedástico para a maioria dos métodos estatísticos, os métodos propostos têm bom desempenho em ambos os cenários. Isto confirma que os métodos de regressão quantílica propostos são robustos à heterocedasticidade. Seu desempenho é testado nos dados do painel econômico da Organização para Cooperação e Desenvolvimento Econômico (OCDE). O objetivo do terceiro capítulo é restringir simultaneamente o número de constantes de regressão individuais e covariáveis explicativas. Além do LASSO, é proposto um LASSO adaptativo que permite a seleção assimptótica do modelo verdadeiro, se este existir, e que desfruta da propriedade de normalidade assimptótica clássica. As simulações de Monte Carlo são realizadas no caso de baixa dimensionalidade (muito mais observações do que parâmetros) e no caso de dimensionalidade moderada (número equivalente de observações e parâmetros). Em ambos os casos, o método adaptativo tem um desempenho muito melhor do que os métodos não adaptativos. Finalmente, aplicamos nossa metodologia em um conjunto de dados de coorte de dimensionalidade moderada. Para cada capítulo, um software de código aberto é escrito e colocado à disposição da comunidade científica
Mahieu, Céline. "Représentations de l'engagement en doctorat des sages-femmes enseignantes." Electronic Thesis or Diss., Normandie, 2023. http://www.theses.fr/2023NORMR018.
Full textIn 2014, 9,1 % of teaching midwives and midwifery school directors held a doctoral degree or were in the process of doctoral training in France (Morin & Leymarie, 2016). However, only a master's degree is required for such professional assignments. We therefore wondered why a significant proportion of midwifery teachers are enrolled in doctoral studies. Then we wanted to find out how they manage to stay committed during their doctoral training years when they are returning to their studies with an already established private and professional life.To answer this question, we studied the context of this phenomenon and the concepts related to our research question such as commitment in training (De Ketele, 2013b; Kaddouri, 2011), the doctoral training process (Cros & Bombaron, 2018; Skakni, 2019), the life course (Sapin and al., 2014) and gender and care (Molinier and al., 2009; Paperman, 2013; Champagne and al., 2015). In addition, we conducted an empirical study with a qualitative and longitudinal methodology including two rounds of interviews at one-year intervals with midwifery teachers in doctoral training at least in the first interview in spring 2021. Themes were coded using NVivo software, followed by a longitudinal and cross-sectional content analysis (Bardin, 1989).Our results show that the process of universitarisation of initial midwifery education in France, which has been underway since 2009, is an important motivational source for midwifery teachers' commitment to a doctorate. However, midwifery teachers involved in doctoral studies also have a personal intellectual motivation for reflective analysis of professional midwifery and midwifery teaching practices, often in a quest for recognition of their medical professional identity.Their commitment is strong at the behavioural, cognitive and emotional levels, to use the indicators of commitment to training according to Pintrich et al (1993). Various factors influence their commitment to the doctorate, notably the articulation between their personal, professional and doctoral life trajectories, the recognition of their doctoral work by their hierarchy, and the relationship with their thesis director. In short, midwifery teachers consider the commitment to a doctorate as a respectable value and as an investment for themselves, for the midwifery profession and for the midwifery science.This research has thus highlighted the phenomenon of doctoral commitment in one of the care professions in the current context of universitarisation. It also highlights the difficulties encountered by a population of women returning to study for a doctorate. Certain levers are highlighted in the scientific literature and through the results of our empirical study. Research perspectives emanate from this thesis in order to find ways to improve the quality of doctoral life in terms of well-being and academic performance
Maillard, Angéline. "Atteintes cognitives et cérébrales dans le trouble de l'usage d'alcool et le syndrome de Korsakoff : valeur pronostique, évolution et prise en charge Prognosis factors of low-risk drinking and relapse in alcohol use disorder : a multimodal analysis Short-term neuropsychological recovery in alcohol use disorder : a retrospective clinical study Is there cognitive and brain changes over time in Korsakoff's syndrome ? Neuropsychological deficits in alcohol use disorder : impact on treatmen The effect of alcohol withdrawal syndrome severity on sleep, brain and cognition." Thesis, Normandie, 2020. http://www.theses.fr/2020NORMC017.
Full textAlcohol use disorder (AUD) is characterized by brain damage and cognitive deficits. These alterations hinder AUD patients to benefit from psychosocial treatment and increase the risk of relapse. It is now clear that cognitive deficits and brain abnormalities can be reversible with drinking cessation in AUD. However, patients with Korsakoff’s syndrome (KS) are described as exhibiting a severe anterograde amnesia supposed to persist over time, even though longitudinal studies in KS patients are very rare. The objective of this thesis is to examine the prognostic value, changes over time, and rehabilitation of the cognitive impairments and brain alterations in AUD and KS. Our results suggest that alexithymia, as well as alteration of limbic and frontocerebellar systems observed early in abstinence, contribute to a poor prognosis regarding alcohol status within the year following detoxification. We highlight that, after detoxification, a short stay as inpatient in a convalescent home favors cognitive improvement, and even a return to a normal level of performance. During this stay, an intensive care including neuropsychological training seems to favor the recovery. Finally, our results indicate that in KS patients, severe memory impairments, sustained by Papez circuit alterations, persist over time. Executive deficits and damage of the fronto-cerebellar circuit may recover but to a limited extent. These results emphasize the need to assess cognitive and brain alteration that have a prognostic value regarding treatment outcome. Results also encourage adapting treatment to favor recovery in AUD, or to compensate for persisting memory impairments in KS
Abichou, Klich Amna. "Décomposition de la variance dans le modèle de classification de trajectoires de biomarqueurs." Thesis, Lyon, 2019. http://www.theses.fr/2019LYSE1199/document.
Full textThe analysis of longitudinal measures –called trajectories– is more and more frequent in clinical research. One of the interests of this analysis is to identify groups of individuals with similar trajectories. The obtained classification is used to understand and explore the heterogeneity of trajectories among subjects. The classification can be performed by a model that predicts the same trajectory for all the subjects that are classified in the same group. The objective of this thesis is to develop an extension to the standard classification model that gives greater consideration to the variability within groups, (i) the variability of marker values (residual variance), and (ii) the variability of the individual trajectories inside a group (between-individual variance). Two classification models were developed: 1) a first model that allows unequal residual variance across groups, and 2) a second model that takes into account a between-individual variance within each group instead of predicting the same trajectory for all subjects in the same group, a variance that can be equal or unequal across groups. The interest of these two models has been studied by simulations and through clinical applications. Overall, when the number of trajectories and measurements per trajectory is sufficient, these models gives better classification compared to the standard classification model. Moreover, except for highly controlled experimental designs, the two sources of variability are inherent to research in health. Therefore, these models are very relevant from a clinical point of view
Ortholand, Juliette. "Joint modelling of events and repeated observations : an application to the progression of Amyotrophic Lateral Sclerosis." Electronic Thesis or Diss., Sorbonne université, 2024. http://www.theses.fr/2024SORUS227.
Full textProgression heterogeneity in chronic diseases such as Amyotrophic Lateral Sclerosis (ALS) is a significant obstacle to developing effective treatments. Leveraging the growing wealth of large databases through modelling can help better understanding it. However, the data collected only offer access to partial trajectories, that need to be realigned to reconstruct a comprehensive disease progression. To address this challenge, data-driven progression models like the longitudinal Spatiotemporal model were developed. Its main interest is its ability to synchronise patients onto a common disease timeline (temporal aspect) thanks to a latent disease age, while also capturing the remaining variability through parameters that account for outcome ordering (spatial aspect). However, this model was primarily designed for longitudinal data, overlooking crucial outcomes in ALS such as time to death or initiation of life support, like Non-Invasive Ventilation (NIV). Conversely, existing joint models offer the advantage of simultaneously handling longitudinal and survival data. However, they do not realign trajectories, which compromises their temporal resolution. This thesis aimed to expand the Spatiotemporal model into a Joint Spatiotemporal model, enabling, for ALS research, the examination of survival data alongside longitudinal data. First, we applied the Spatiotemporal model to explore how the interaction between sex and onset site (spinal or bulbar) impacts the progression of ALS patients. We selected 1,438 patients from the PRO-ACT database. We demonstrated a significant influence of both sex and onset site on six longitudinal outcomes monitoring the functional and respiratory decline in addition to Body Mass Index. However, this study did not incorporate survival analysis, despite its paramount importance in ALS, due to limitations inherent to the Spatiotemporal model. To address this gap, we associated the Spatiotemporal model with a survival model that estimates a Weibull survival model from its latent disease age, creating a univariate Joint Temporal model. After model validation, we benchmarked our model with a state-of-the-art joint model on PRO-ACT data. Our model exhibited significantly superior performance in terms of absolute bias and mean cumulative AUC for right-censored events. This demonstrated the efficacy of our approach in the context of ALS compared to existing joint models. However, modelling several longitudinal outcomes requires a multivariate approach. Life support initiation that might be censored by death needs to be also considered. We thus extended the Joint Temporal model, into a multivariate Joint Spatiotemporal model with competing risks to analyse NIV initiation. This involved coupling the multivariate Spatiotemporal model with a cause-specific Weibull survival model from the latent disease age. We incorporated spatial parameters with a Cox proportional effect on the hazard. After validation, we benchmarked our model with a state-of-the-art joint model on PRO-ACT data and analysed sex and onset site interaction in complement to the first study. The Joint Spatiotemporal model achieved similar performance to the state-of-the-art model while capturing an underlying shared latent process, the latent disease age, whereas the state-of-the-art models the impact of longitudinal outcomes on survival. To enhance the reproducibility and facilitate the reuse of these models, the proposed models were implemented in the open-source software Leaspy. In conclusion, this thesis introduces the first data-driven progression model combining longitudinal and survival modelling. We demonstrated its relevance to understand the occurrence of critical events in ALS. This work paves the way for further extension to analyse recurrent events, among other potential applications in causal inference
Barhoumi, Mohamed Adel. "Traitement des données manquantes dans les données de panel : cas des variables dépendantes dichotomiques." Thesis, Université Laval, 2006. http://www.theses.ulaval.ca/2006/23619/23619.pdf.
Full textGafsi, Nicolas. "Relations entre dynamique de production laitière, gestion des réserves corporelles et performances de reproduction chez la chèvre laitière : une approche par modélisation des trajectoires phénotypiques." Electronic Thesis or Diss., université Paris-Saclay, 2024. http://www.theses.fr/2024UPASB027.
Full textIn a context of agro-ecological transition and increasingly frequent climatic hazards, goat systems in France are facing several technical and socio-economic challenges. Goat farming systems need to breed robust goats capable of producing and reproducing in potentially disturbed environments. However, in France, the average productive lifetime of a goat is short. Implementing herd management strategies and genetic selection to improve the longevity and robustness of dairy goats imply to consider the interactions between the different biological functions and their dynamic change at the lactation and lifetime scales. This multivariate and dynamic view is crucial to understand robustness and diversity of lifetime productive trajectories. The objective of this thesis is to use a modelling approach of phenotypic trajectories of milk yield (MY), body condition score (BCS) and body weight (BW) to understand the relationships between lactation, body reserves and reproduction. Data from two goat experimental stations in France collected between 1996 and 2020 were used to characterize the diversity of phenotypic curves of MY, BW, and BCS throughout lactation and explore the associations between these different phenotypic curves, reflecting lactation and body reserves management. Lifetime scale was studied by analyzing the sequence of phenotypic trajectories over parities and the relationships with reproductive success, in particular for artificial insemination (AI). Finally, we used an existing herd demography model to put into perspective and assess the consequences of the diversity of phenotypic trajectories at the herd scale and the corresponding reproductive performance. Results showed a diversity of phenotypic trajectories for MY, BW, and BCS at the lactation level and no strong associations between the different trajectories, suggesting a diversity of strategies for energy partitions between the different biological functions. Over several lactations, there was a repeatability of phenotypic trajectories membership across parities. Regarding reproductive performance, goats with the highest MY persistency were significantly less likely to succeed at AI than goats with low MY persistency. For BW, only the dynamics of repletion around AI influenced success at AI. For BCS, it was the minimum level reached during lactation that influenced AI success. Simulation approaches showed that a drop in fertility at AI had no major effect on the distribution of kidding. This work proposes an original methodological approach for phenotypic time-series data and opens perspective to provide technical references for feeding and reproductive management of dairy goat herds
Couronné, Raphaël. "Modélisation de la progression de la maladie de Parkinson." Electronic Thesis or Diss., Sorbonne université, 2021. http://www.theses.fr/2021SORUS363.
Full textIn this work, we developed statistical methods to model disease progression from patient’s repeated measurements, with a focus on Parkinson’s Disease (PD). A key challenge lies in the inherent heterogeneity of PD across patients, to the extent that PD is now suspected to encompass multiple subtypes or motor phenotypes. To gain insights on disease progression, research studies propose to gather a broad range of marker measurements, at multiple timepoints for each patients. These data allow to investigate the disease’s patterns of progression via statistical modeling. In a first part, we modeled the progression of scalar markers of PD. We extended on a disease progression model, namely the longitudinal spatiotemporal model. We then proposed to address data missingness, and to model the joint progression of markers of different nature, such as clinical scores, and scalar measurements extracted on imaging modalities. With this method, we modeled early motor progression in PD, and, in a second work, the heterogeneity of idiopathic PD progression, with a focus on sleep symptoms. In a second, independent, part of the manuscript, we tackled the longitudinal modeling of medical images. For these higher dimensionality data, Deep Learning is often used, but mostly in cross sectional setups, ignoring the possible inner dynamics. We proposed to leverage Deep Learning as a dimensionality reduction tool to build a spatiotemporal coordinate system of disease progression. We first took advantage of this flexibility to handle multimodal data. Then we leveraged the self-supervision induced by assuming monotonicity over time, to offer higher flexibility in modeling temporal variability
Hombourger-Barès, Sabrina. "La contribution du design de l'espace de vente à l'évolution du positionnement de l'enseigne : une analyse longitudinale." Thesis, Dijon, 2014. http://www.theses.fr/2014DIJOE002/document.
Full textOne of the innovative ways favoured by retailers to drive change in their value proposition is to review the design of their stores. Academic contributions to the in-store experience have mostly focused on consumer perspective and identifying relevant managerial practices. The core of this research studies how repositioning a retail brand translates into the experiential design of retail spaces. To this end, the research follows the repositioning process from a managerial perspective and updates the mechanisms that underlie it. The longitudinal study of embedded cases reveals the importance of an holistic design that takes into account the desired interactions between the shopper and the store. The analysis shows the four overlapping phases of the store’s life cycle, and breaks down the process into six dimensions, each with its own events and issues. The six dimensions are vision, plotline, action, decor, assessment and coproduction.The entrepreneurial vision of the leader is the cornerstone of the whole innovation process. The value proposition is embodied by three components, namely plotline, action and decor. For each of the five stages of the shopper’s journey, elements of the decor are implemented to relay or reinforce the desired action. These are mechanisms or devices meant to stimulate the shopper’s experiential system. The assessment, which involves measuring the perceived and experienced positioning, helps to adjust the value proposition in terms of four levels of consistency and flexibility of design. Finally, the coproduction of store design between different stakeholders can cause a co-destruction of value, whether intentional or accidental
Betoko, Aisha. "L’alimentation précoce : ses déterminants, son influence sur la croissance postnatale et les consommations alimentaires à 3 ans." Thesis, Paris 11, 2013. http://www.theses.fr/2013PA11T038/document.
Full textBackground: Early eating patterns can determine later eating habits and food preferences and they have been related child growth. In the literature, the determinants and health effects of breastfeeding and complementary feeding practices have often been analyzed separately. Yet, breastfeeding and complementary feeding practices are interrelated and there are arguments to suggest that both influence later health. Objectives : We aimed to characterize feeding practices over the first year of life and to examine their associations with family and infant characteristics, with growth changes in the first 3 years of life, and their relations with food intake at 3 years of age. Methodes : Subjects were participants of the EDEN mother-child cohort. The study recruited 2,002 pregnant women aged 18-45 years attending their prenatal visit before 24 weeks’ gestation at Nancy and Poitiers University Hospitals between 2003 and 2006. Dietary practices and anthropometric measurements were collected through maternal self-report and clinical examinations at birth, 4, 8, 12, 24 and 36 months. Principal component analysis was applied to derive patterns from breastfeeding duration, age of introduction of complementary foods (CF) and type of food used at 1y (ready-prepared baby foods, home-made foods, ready-prepared ordinary foods). Multiple linear and logistic regressions were used to analyze associations between feeding patterns, growth and food intake at 3 years of age. Results : i) The type infant formula (partially hydrolyzed, thickened, enriched in pre- or probiotic and others) used in the first four months of life was related to maternal return to employment, parity but not to infant growth in the same period. ii) Three major feeding patterns were identified in the EDEN study. The main source of variability in infant feeding was characterized by a pattern labeled ‘Late weaning and use of ready-prepared baby foods’. Older, more educated, primiparous women with high monthly income and recruited in Nancy ranked high on this pattern. The second pattern, labeled ‘Longer breastfeeding, late CF introduction and use of home-made foods’ was the closest to infant feeding guidelines. Mothers with high scores on this pattern were older, more educated and recruited in Poitiers. The third pattern labeled ‘Use of ordinary foods’ is more suggestive of infants having a less age-specific diet. Mothers ranking high on this pattern were often younger, multiparous and recruited in Nancy. iii) High scores on the second pattern were related to significant lower 0-1y weight and height change, higher 1-3y weight and height change and to a significant higher fruit and vegetables intake at 3 years of age after controlling for a wide range of potential confounding variables. An additional adjustment on breastfeeding duration attenuated the relationships without making them disappear completely, suggesting an effect of the overall feeding practices in the first year of life on the parameters that we studied. Conclusions : Our results confirm the importance of socio-cultural determinants on feeding practices over the first year of life. They also confirm the relations between early nutrition and growth in the first three years of life and later eating habits. Our results emphasize the need to consider infant feeding over the first year of life including breastfeeding duration, age of complementary foods introduction as well as type of foods used when examining effects of early infant feeding practices on later health
Barbieri, Antoine. "Méthodes longitudinales pour l’analyse de la qualité de vie relative à la santé en cancérologie." Thesis, Montpellier, 2016. http://www.theses.fr/2016MONTS026/document.
Full textThe health-related quality of life is a major objective in oncology clinical trials to improve patients’ care and better evaluate the impact of the treatments on their everyday life. Auto-questionnaires are usually used to measure this endpoint. In this work, different statistical models for the longitudinal analysis of health-related quality of life in oncology are proposed and applied to clinical trial data. First, we present different models derived from the item response theory (IRT) to achieve a longitudinal analysis directly on raw data (multi-response outcomes) for each dimension. Within the generalized linear mixed model background, a conceptual selection of the IRT models concluded that the graded response model seems to be the most suitable. Then, we propose a structural equation model which allows taking into account the multidimensional nature of data at each time and the longitudinal aspect induced by the repeated measurements. At each measurement time, the model allows to link all the observed variables issued from the questionnaire given explanatory variables. Two factors are estimated, each summarizing a set of observed variables. The longitudinal analysis is performed on the global health status and on the factors, thus reducing the number of tests. Finally, an approach based on a mixture of mixed models is used to obtain latent classes from quality of life trajectories. The approach has resulted in the identification of homogeneous subpopulations and their latent trajectory according to specific patient profiles
Rouquette, Alexandra. "Mesures subjectives et épidémiologie : problèmes méthodologiques liés à l’utilisation des techniques psychométriques." Thesis, Paris 11, 2014. http://www.theses.fr/2014PA11T097/document.
Full textRecently, subjective measurements have increasingly been used in epidemiology, alongside the growing will to integrate individuals’ point of view on their health in studies on diseases or health interventions. Psychometrics includes statistical methods used to develop questionnaires and to analyze questionnaire data. This doctoral dissertation aimed to explore methodological issues raised by the use of psychometric techniques in epidemiology. Three empirical studies are presented and cover 1 / the validation stage of a questionnaire: the objective was to develop, using simulated data, a tool to determine sample size for internal validity studies on psychiatric scale; 2 / the mathematical properties of the subjective measurement: the objective was to compare the performances of the minimal clinically important difference of a questionnaire, assessed on data from a cohort study, computed using the classical test theory (CTT) framework or the item response theory framework (IRT); 3 / its use in a longitudinal design: the objective was to compare, using simulated data, the performances of a statistical method aimed to analyze the longitudinal course of a subjective phenomenon measured using the CTT or IRT framework, especially when some of the available items used for its measurement differ at each time of data collection. Finally, directed acyclic graphs were used to discuss the results from these three studies and the concept of information bias when subjective measurements are used in epidemiology
Nicolaï, Alice. "Interpretable representations of human biosignals for individual longitudinal follow-up : application to postural control follow-up in medical consultation." Electronic Thesis or Diss., Université Paris Cité, 2021. http://www.theses.fr/2021UNIP5224.
Full textIndividual longitudinal follow-up, which aims at following the evolution of an individual state in time, is at the heart of numerous public health issues, particularly in the field of medical prevention. The increasing availability of non-invasive sensors that record various biosignals (e.g., blood glucose, heart rate, eye movements), has encouraged the quantification of human physiology, sensorimotricity, or behavior with the purpose of deriving markers for individual follow-up. This objective raises however several challenges related to signal modelling. Indeed, this particular type of data is complex to interpret, and, a fortiori, to compare across time. This thesis studies the issue of extracting interpretable representations from biosignals through the problematic of balance control follow-up in medical consultation, which has crucial implications for the prevention of falls and frailty in older adults. We focus in particular on the use of force platforms, which are commonly used to record posturography measures, and can be easily deployed in the clinical setting thanks to the development of low cost platforms such as the Wii Balance Board. For this particular application, we investigate the pros and cons of using feature extraction methods or alternatively searching for a generative model of the trajectories. Our contributions include first the review and study of a wide range of state-of-the-art variables that are used to assess fall risk in older adults, derived from the center of pressure (CoP) trajectory. This signal is commonly analyzed in the clinical literature to infer information about balance control. Secondly, we develop a new generative model, ``Total Recall'', based on a previous stochastic model of the CoP, which has shown to reproduce several characteristics of the trajectories but does not integrate the dynamic between the CoP and the center of mass (CoM) -- a dynamic which is considered to be central in postural control. We also review and compare the main methods of estimation of the CoM in quiet standing and conclude that it is possible to obtain an accurate estimation using the Wii Balance Board. The results show the potential relevance of the Total Recall model for the longitudinal follow-up of postural control in a clinical setting. Overall, we highlight the benefit of using generative models, while pointing out the complementarity of features-based and generative-based approachs. Furthermore, this thesis is interested in introducing representations learned on labeled data and tailored for a particular objective of follow-up. We propose new classification algorithms that take advantage of a priori knowledge to improve performances while maintaining complete interpretability. Our approach relies on bagging-based algorithms that are intrinsically interpretable, and a model-space regularization based on medical heuristics. The method is applied to the quantification of fall risk and frailty. This dissertation argues for the importance of researching interpretable methods, designed for specific applications, and incorporating a-priori based on expert knowledge. This approach shows positive results for the integration of the selected biosignals and statistical learning methods in the longitudinal follow-up of postural control. The results encourage the continuation of this work, the further development of the methods, especially in the context of other types of follow-up such as continuous monitoring, and the extension to the study of new biosignals
Canouil, Mickaël. "Développement et application de méthodologies statistiques pour études multi-omiques dans le diabète de type 2 : au-delà de l'ère des études d'association pangénomiques." Thesis, Lille 2, 2017. http://www.theses.fr/2017LIL2S017/document.
Full textGenome-wide association studies (GWAS) have resulted in the identification of several dozen of genes and single nucleotide polymorphisms (SNPs) contributing to type 2 diabetes.More generally, GWAS have identified thousands of SNPs contributing to complex diseases in humans.However, the functional characterization and biological mechanisms involving these SNPs and genes remain largely to be explored. Indeed, the consequences of these polymorphisms are complex and little known.One direct consequence of the SNPs is the alteration of the protein encoded by a gene, or even a complete transcriptional gene silencing (e.g. codon stop in the sequence). Furthermore, these polymorphisms may have a regulatory role in gene expression, for example, by interfering with the binding of transcription factors and enzymes involved in DNA methylation.Despite the strong associations of identified SNPs, they cannot explain the full heritability of type 2 diabetes, suggesting interactions mechanisms between the different layers of -omics, such as genomics, transcriptomics and Epigenomics.The shift of paradigm in statistical genetics and the availability of transcriptomic and epigenomic data are responsible for the evolution of the discipline, moving from association studies to multi-omics, and providing insights on the functional aspect of the SNPs or genes involved, and in some cases allowing to evaluate the causal link of these variants on the pathology.The methodological developments and their applications proposed in this thesis are various, ranging from a similar approach to GWAS, leveraging the longitudinal data available in some cohorts (e.g. D.E.S.I.R.), using an joint model approach; the functional characterisation of candidate genes in insulin secretion by a multi tissue transcriptomic study and transcriptomic study in a cell model; the identification of a new candidate gene (PDGFA) involved in the deregulation of the insulin\\\'s pathway in type 2 diabetes through epigenetic and transcriptomic mechanisms; and finally, the characterisation of the effect on the transcriptome of two substitutes of bisphenol A in a primary adipocyte model.The increase of knowledge in biological processes involving SNPs and genes identified by GWAS could enable the development of more effective diagnostic strategies, and the identification of therapeutic targets for the treatment of type 2 diabetes and associated complications (e.g., insulin resistance, NAFLD, cancer, etc.).More generally, these multi-omics studies pave the way for the emerging approach of precision medicine, allowing the treatment and prevention of pathologies while taking into account what makes the specificity of an individual, namely his genome and his environment, both interacting on his transcriptome and his epigenome
Loucoubar, Cheikh. "Statistical genetic analysis of infectious disease (malaria) phenotypes from a longitudinal study in a population with significant familial relationships." Phd thesis, Université René Descartes - Paris V, 2012. http://tel.archives-ouvertes.fr/tel-00685104.
Full textJia, Dawei. "Évaluation des contraintes internes par méthode ultrasonore dans les pièces plastiques injectées." Thesis, Lille 1, 2009. http://www.theses.fr/2009LIL10107.
Full textThe mechanical properties and the dimensional precision of the injection molded polymer parts are influenced by the process-induced residual stresses. This study shows the potential of the ultrasonic method based acoustoelastic effect to evaluate the residual stresses in amorphous polymer plates with Lcr waves. Firstly, special holding fixtures were designed according to their function analyses and the optimization of some parameters (ultrasonic field, incidence angle and emitter-receiver distance). They were used to determine the acoustoelastic coefficients K33 and K11, as well as the coefficient P corresponding to the temperature effect on the longitudinal wave velocity in studied materials. Then ultrasonic measurements were applied to partially relaxed PC and PS plates, through C-Scan system to evaluate average stresses across plate thickness, and Lcr waves to evaluate the surface stress. The measurement reliability was validated by comparing the results with those of the photoelasticity. The poor plate thickness (3 mm) affects the measure accuracy of the speed scanning, while it is improved a lot by using Lcr waves (Ds » ±2 MPa), so we turn to the latter only. The last part of the study concerns the determination of the stresses after manufacturing of new plates and that of stress profile in different thicknesses by changing the frequency of the wave Lcr
Juillard, Hélène. "Méthodes d'estimation et d'estimation de variance pour une enquête longitudinale : application aux données de l'Etude Longitudinale Française depuis l'Enfance (Elfe)." Thesis, Toulouse 1, 2016. http://www.theses.fr/2016TOU10026/document.
Full textIn this document, we are interested in estimation under a design-based framework, where the randomness arises from the sample selection. Each sampling leads to a sampling variance. After the survey, the estimation of this variance will serve as a measure of precision (or uncertainty) for the estimators of the parameters under study. The 2011 ELFE cohort comprises more than 18,000 children whose parents consented to their inclusion. In each of the selected maternity units, targeted babies born during four specific periods representing each of the four seasons in 2011 were selected. ELFE is the first longitudinal study of its kind in France, tracking children from birth to adulthood. It will examine every aspect of these children’s lives from the perspectives of health, social sciences and environmental health. The ELFE cohort was selected through a non-standard sampling design that is called cross-classified sampling, with independent selections of the sample of maternity units and of the sample of days. In this work, we propose unbiased variance estimators to handle this type of sampling designs, and we derive specific variance estimators adapted to the ELFE case. Tracking of the babies starts when they are just a few days old and still at the maternity unit. When the children reach the age of two months, the parents are contacted for the first telephone interview. When the children are one year old, and again when they reach the ages of two, three and a half years and five and a half years, their parents will once more be contacted by telephone. The survey is longitudinal.The first chapter of this thesis introduces concepts related to the theory of survey design and presents the survey ELFE (French Longitudinal Study from Childhood); its data will be used as illustration for the theoretical results derived in this thesis. The second chapter focuses on the cross-classified design and provides unbiased estimators and simplified variance estimators to treat this design in a general theoretical framework. It is also shown that this design is generally less efficient than the conventional two-stage sampling design. Chapter three is in continuity with the previous one : for the cross-classified sampling design, five unbiased Yates-Grundy like variance estimators are available from five different possible decomposition of the variance. Chapter four is an article allowing the reader to make the difference between the cross-classified sampling design and the two-stage sampling design, and to implement the steps of sampling and estimation under the softwares R, SAS and Stata. Chapter five is devoted to variance computation and variance estimation for a cohort survey with monotone non-response. Chapter six is a methodological report to users in which the appropriate variance estimation for the ELFE design is explained and implemented with softwares R, SAS and Stata. All the results of simulation studies presented in this document are reproducible, the codes being proposed in the annex
Kocevar, Gabriel. "Développement de méthodes d’IRM avancées pour l’étude longitudinale de la Sclérose en Plaques." Thesis, Lyon, 2017. http://www.theses.fr/2017LYSE1057/document.
Full textWhile conventional MRI is the reference tool for the diagnosis and monitoring of MS, it remains only moderately correlated with the patient’s clinical status. In order to better characterize pathological alterations occurring in MS, we use in this work non-conventional MRI techniques, namely magnetic resonance spectroscopy (MRS) and diffusion MRI.A first weekly follow-up revealed the sensitivity of the diffusion metrics and the specificity of the SRM to detect the initial processes of lesion formation.A second follow-up revealed changes in diffusivity in several white matter fiber bundles, including a decrease in fraction of anisotropy and an increase in radial diffusivity, worsening with advancing disease and more marked in the progressive forms.Finally, the application of graph theory allowed to characterize the brain connectivity in the four clinical forms and to study their evolution. This study allowed us to highlight alterations in all the four clinical phenotypes, including a decrease in the cerebral network density, more marked in the progressive forms of the disease and tending to increase with its progression.This work shows the sensitivity of advanced MRI techniques for the characterization of pathological alterations and their evolution in MS
Fiot, Jean-Baptiste. "Méthodes mathématiques d’analyse d’image pour les études de population transversales et longitudinales." Thesis, Paris 9, 2013. http://www.theses.fr/2013PA090053/document.
Full textIn medicine, large scale population analysis aim to obtain statistical information in order to understand better diseases, identify their risk factors, develop preventive and curative treatments and improve the quality of life of the patients.In this thesis, we first introduce the medical context of Alzheimer’s disease, recall some concepts of statistical learning and the challenges that typically occurwhen applied in medical imaging. The second part focus on cross-sectional studies,i.e. at a single time point. We present an efficient method to classify white matter lesions based on support vector machines. Then we discuss the use of manifoldlearning techniques for image and shape analysis. Finally, we present extensions ofLaplacian eigenmaps to improve the low-dimension representations of patients usingthe combination of imaging and clinical data. The third part focus on longitudinalstudies, i.e. between several time points. We quantify the hippocampus deformations of patients via the large deformation diffeomorphic metric mapping frameworkto build disease progression classifiers. We introduce novel strategies and spatialregularizations for the classification and identification of biomarkers
Fiot, Jean-Baptiste. "Méthodes mathématiques d'analyse d'image pour les études de population transversales et longitudinales." Phd thesis, Université Paris Dauphine - Paris IX, 2013. http://tel.archives-ouvertes.fr/tel-00952079.
Full textTacchella, Jean-Marc. "Méthodes d'analyse de volumes d'images multimodales pour le suivi longitudinal en oncologie." Thesis, Paris 6, 2015. http://www.theses.fr/2015PA066049/document.
Full textLongitudinal follow-up in oncology consists in assessing tumor progression in order to define a treatment strategy for each patient. It is thus necessary to identify relevant biomarkers that allow an early evaluation of the tumor’s response to the treatment. Biomedical imaging exams help to ensure a non-invasive monitoring of patients and to propose complementary biomarkers to existing ones. This search for new biomarkers requires relying on clinical studies to compare globally and locally exams acquired at different times and in different modalities.In this PhD dissertation, we focused on developing an integrated image processing framework to analyze the relevance of indices for evaluating the tumor evolution. The strategy includes three steps: registration data acquired at different times and in different modalities, the segmentation of tumor lesions on exams in each modality, and the computation of global and local indices reflecting the spatiotemporal evolution of the tumor.The most innovative aspect lies in the registration step: due to the difficulties faced with conventional methods, we proposed a new approach consisting in using several registration methods and selecting the best one for each dataset, thanks to a quantitative criterion based on the specific features of the application. The contribution of this approach was proven in two clinical studies: 1) monitoring of patients with high-grade gliomas treated with an antiangiogenic drug, where Single Photon Emission Tomography data (SPECT) obtained after injection of Technetium-99m Sestamibi have to be matched with T1-weighted Magnetic Resonance Images (MRI) acquired after the injection of a contrast agent; 2) monitoring of patients with liver damage treated with various anticancer drugs, requiring the alignment of Computed Tomography (CT) data.The complete image processing framework was applied to the first study. Tumor areas were segmented on MR images using a conventional 2D Level Set method, as well as on the TEMP data using five thresholding methods that differ in the choice of the threshold options. Despite a strong correlation in terms of overall volumes, local indices have shown that some of the detected tumor volumes on early SPECT exams (performed 15 minutes after injection) and late SPECT exams (performed 3 hours after injection) could be located out of the tumor volumes detected on MRI. The high correlation found between the intensity variations in tumor volume during treatment on late SPECT exams and the index of Overall Survival (OS) suggests that this relative change of intensity could be predictive of the patient overall survival, which is not the case with the indices derived from the MRI data on our limited series of patients. Thus, these results show that SPECT imaging, with an exam performed 3 hours after injection of Technetium-99m Sestamibi, can be complementary to MRI in the assessment of tumor progression in glioblastomas.The main perspective for this PhD work would consist in applying the analysis strategy to other clinical studies. However, each step must be adapted to suit the specific nature of the targeted application, including the imaging modalities involved and the considered anatomical area. The expected sticking points are the automation and the robustness of the different steps of the processing chain
Tacchella, Jean-Marc. "Méthodes d'analyse de volumes d'images multimodales pour le suivi longitudinal en oncologie." Electronic Thesis or Diss., Paris 6, 2015. http://www.theses.fr/2015PA066049.
Full textLongitudinal follow-up in oncology consists in assessing tumor progression in order to define a treatment strategy for each patient. It is thus necessary to identify relevant biomarkers that allow an early evaluation of the tumor’s response to the treatment. Biomedical imaging exams help to ensure a non-invasive monitoring of patients and to propose complementary biomarkers to existing ones. This search for new biomarkers requires relying on clinical studies to compare globally and locally exams acquired at different times and in different modalities.In this PhD dissertation, we focused on developing an integrated image processing framework to analyze the relevance of indices for evaluating the tumor evolution. The strategy includes three steps: registration data acquired at different times and in different modalities, the segmentation of tumor lesions on exams in each modality, and the computation of global and local indices reflecting the spatiotemporal evolution of the tumor.The most innovative aspect lies in the registration step: due to the difficulties faced with conventional methods, we proposed a new approach consisting in using several registration methods and selecting the best one for each dataset, thanks to a quantitative criterion based on the specific features of the application. The contribution of this approach was proven in two clinical studies: 1) monitoring of patients with high-grade gliomas treated with an antiangiogenic drug, where Single Photon Emission Tomography data (SPECT) obtained after injection of Technetium-99m Sestamibi have to be matched with T1-weighted Magnetic Resonance Images (MRI) acquired after the injection of a contrast agent; 2) monitoring of patients with liver damage treated with various anticancer drugs, requiring the alignment of Computed Tomography (CT) data.The complete image processing framework was applied to the first study. Tumor areas were segmented on MR images using a conventional 2D Level Set method, as well as on the TEMP data using five thresholding methods that differ in the choice of the threshold options. Despite a strong correlation in terms of overall volumes, local indices have shown that some of the detected tumor volumes on early SPECT exams (performed 15 minutes after injection) and late SPECT exams (performed 3 hours after injection) could be located out of the tumor volumes detected on MRI. The high correlation found between the intensity variations in tumor volume during treatment on late SPECT exams and the index of Overall Survival (OS) suggests that this relative change of intensity could be predictive of the patient overall survival, which is not the case with the indices derived from the MRI data on our limited series of patients. Thus, these results show that SPECT imaging, with an exam performed 3 hours after injection of Technetium-99m Sestamibi, can be complementary to MRI in the assessment of tumor progression in glioblastomas.The main perspective for this PhD work would consist in applying the analysis strategy to other clinical studies. However, each step must be adapted to suit the specific nature of the targeted application, including the imaging modalities involved and the considered anatomical area. The expected sticking points are the automation and the robustness of the different steps of the processing chain
Legouhy, Antoine. "Méthodes pour la modélisation morphologique longitudinale du cerveau via le recalage et la création d'atlas." Thesis, Rennes 1, 2020. http://www.theses.fr/2020REN1S008.
Full textUnderstanding brain development involves studying the relationship between age as one of the explanatory variables and explained variables, observations of this organ, which can take many forms. Magnetic Resonance Imaging (MRI) gives the opportunity to extract such observations in a non-invasive and non-irradiating way. This powerful technique allows notably to gain insights about the functional activity of the brain or its internal diffusivity characteristics. Yet, it is rather on the purely morphological aspects that this thesis is focused on. The approach followed the study of the brain as a mathematical object, thus enabling the analysis of its shape and growth by the means of the geometric transformations connecting those objects. In the finding of those transformations, across structures of topological interest, lies the concept of registration. This opens the door to the statistical analysis of shapes and the creation of average anatomical models called atlases
Hanan, Audrey. "L'influence de la transgression des normes de distribution de l'enseigne sur son image de marque et sa relation avec le consommateur : le cas des produits alimentaires dits "moches"." Thesis, Aix-Marseille, 2018. http://www.theses.fr/2018AIXM0607.
Full textThis research has based itself on the relationship between the consumer and the brand to demonstrate the consequences of an act of transgression done by the brand. Such breaches constitute ruptures with customs so widespread that they escape the consumer’s attention, while simultaneously remaining predominant and visible enough to impregnate its consumption experience and ultimately its expectations. This theoretical approach is highlighted by a real case of a breach of distribution norms: the recent offer of “ugly” products by supermarkets, which constitutes a breach to the usual norm of standardised products. This transgression occurred subsequently after the controversy over food waste. The research model presented herein is based on a review of literature, in addition to two qualitative exploratory studies. A longitudinal study examining regular customers of E.Leclerc supermarkets, which was conducted on the same respondents during two different sessions, is used to measure and validate our research hypothesis. The results exhibit the evolution of the consumer’s behavior following the introduction of the transgressive offer and confirm the deterioration of the brand image and the engagement process. If the transgression can be destructive, it can also create value through its innovative nature, its ability to attract consumers and its power of differentiation
Leone, Nathalie. "Fonction ventilatoire, asthme et facteurs de risque cardiométabolique." Phd thesis, Université Paris Sud - Paris XI, 2014. http://tel.archives-ouvertes.fr/tel-00965432.
Full textAltzerinakou, Maria Athina. "Méthodes statistiques pour les essais de phase I/II de thérapies moléculaires ciblées en cancérologie." Thesis, Université Paris-Saclay (ComUE), 2018. http://www.theses.fr/2018SACLS375/document.
Full textConventional dose-finding approaches in oncology of phase I clinical trials aim to identify the optimal dose (OD) defined as the maximum tolerated dose (MTD), based on the toxicity events observed during the first treatment cycle. The constant development of molecularly targeted agents (MTAs), usually administered in chronic schedules, has challenged this objective. Not only, the outcomes after the first cycle are of importance, but also activity does not necessarily increase monotonically with dose. Therefore, both toxicity and activity should be considered for the identification of the OD. Lately, continuous biomarkers are used more and more to monitor activity. The aim of this thesis was to propose and evaluate adaptive designs for the identification of the OD. We developed two dose-finding designs, based on a joint modeling of longitudinal continuous biomarker activity measurements and time to first dose limiting toxicity (DLT), with a shared random effect, using skewed normal distribution properties. Estimation relied on likelihood that did not require approximation, an important property in the context of small sample sizes, typical of phase I/II trials. We addressed the important case of missing at random data that stem from unacceptable toxicity, lack of activity and rapid deterioration of phase I patients. The MTD was associated to some cumulative risk of DLT over a predefined number of treatment cycles. The OD was defined as the lowest dose within a range of active doses, under the constraint of not exceeding the MTD. The second design extended this approach for cases of a dose-activity relationship that could reach a plateau. A change point model was implemented. The performance of the approaches was evaluated through simulation studies, investigating a wide range of scenarios and various degrees of data misspecification. As a last part, we performed an analysis of 27 phase I studies of MTAs, as monotherapy, conducted by the National Cancer Institut. The primary focus was to estimate the per-cycle risk and the cumulative incidence function of severe toxicity, over up to six cycles. Analyses were performed separately for different dose subgroups, as well as for hematologic and non-hematologic toxicities
Paris, Nicolas. "Formalisation algorithmique des classements au tennis : mise en perspective longitudinale par simulation probabiliste." Bordeaux 2, 2008. http://www.theses.fr/2008BOR21603.
Full textBellec, Pierre. "Etude longitudinale des réseaux cérébraux à large échelle en IRMf : méthodes et application à l'étude de l'apprentissage moteur." Paris 11, 2006. http://www.theses.fr/2006PA112011.
Full textSkill learning in human healthy volunteers is thought to induce a reorganization of cerebral activity. Such process of cerebral plasticity involves the modulation of functional interactions within networks of spatially distributed brain regions, or large-scale networks. Various measures of connectivity exist that allow one to quantify these functional interactions using functional magnetic resonance imaging (fMRI), which enables the non-invasive, yet indirect, measure of cerebral activity. I developed a series of methods that allows to characterize the reorganization of large-scale networks in the brain when considering an fMRI longitudinal study of a single subject or group of subjects, at various stages of a learning process. First, the regions of the network involved in the execution of the task under scrutiny are built and identified from the functional data in an exploratory way, by using a competitive region growing method, which segments the gray matter into functionally homogeneous regions, then followed by a statistical classification procedure. A statistical method is then designed to assess which interactions are significantly modulated within the network during the plasticity process. This method is based on a non-parametric bootstrap technique, taking the temporal auto-correlation of fMRI time series into account, and controling the false discovery rate. These methods have been validated and evaluated on both synthetic and real datasets. Two real datasets were studied, which involved learning of a sensorimotor adaptation task and of a motor sequence task, respectively
Carrière, Isabelle. "Comparaisons des méthodes d'analyse des données binaires ou ordinales corrélées. Application à l'étude longitudinale de l'incapacité des personnes âgées." Phd thesis, Université Paris Sud - Paris XI, 2005. http://tel.archives-ouvertes.fr/tel-00107384.
Full textimportant en épidémiologie. L'étude longitudinale de l'incapacité des personnes âgées et la
recherche des facteurs de risque de la vie en incapacité représente un enjeu crucial de santé
publique. Dans ce contexte nous comparons les modèles logistiques marginaux et les modèles
à effets aléatoires en prenant comme réponse l'incapacité considérée comme variable binaire
afin d'illustrer les aspects suivants : choix de la structure de covariance, importance de
données manquantes et des covariables dépendantes du temps, interprétation des résultats. Le
modèle à effets aléatoires est utilisé pour construire un score prédictif de l'incapacité issu
d'une large analyse des facteurs de risque disponibles dans la cohorte Epidos. Les modèles
logistiques ordonnés mixtes sont ensuite décrits et comparés et nous montrons comment ils
permettent la recherche d'effets différenciés des facteurs sur les stades d'incapacité.
Carrière, Isabelle. "Comparaisons des méthodes d'analyse des données binaires ou ordinales corrélées : Application à l'étude longitudinale de l'incapacité des personnes agées." Paris 11, 2005. http://www.theses.fr/2005PA11T030.
Full text