Teses / dissertações sobre o tema "Chaînes de Markov branchantes"
Crie uma referência precisa em APA, MLA, Chicago, Harvard, e outros estilos
Veja os 50 melhores trabalhos (teses / dissertações) para estudos sobre o assunto "Chaînes de Markov branchantes".
Ao lado de cada fonte na lista de referências, há um botão "Adicionar à bibliografia". Clique e geraremos automaticamente a citação bibliográfica do trabalho escolhido no estilo de citação de que você precisa: APA, MLA, Harvard, Chicago, Vancouver, etc.
Você também pode baixar o texto completo da publicação científica em formato .pdf e ler o resumo do trabalho online se estiver presente nos metadados.
Veja as teses / dissertações das mais diversas áreas científicas e compile uma bibliografia correta.
Weibel, Julien. "Graphons de probabilités, limites de graphes pondérés aléatoires et chaînes de Markov branchantes cachées". Electronic Thesis or Diss., Orléans, 2024. http://www.theses.fr/2024ORLE1031.
Texto completo da fonteGraphs are mathematical objects used to model all kinds of networks, such as electrical networks, communication networks, and social networks. Formally, a graph consists of a set of vertices and a set of edges connecting pairs of vertices. The vertices represent, for example, individuals, while the edges represent the interactions between these individuals. In the case of a weighted graph, each edge has a weight or a decoration that can model a distance, an interaction intensity, or a resistance. Modeling real-world networks often involves large graphs with a large number of vertices and edges.The first part of this thesis is dedicated to introducing and studying the properties of the limit objects of large weighted graphs : probability-graphons. These objects are a generalization of graphons introduced and studied by Lovász and his co-authors in the case of unweighted graphs. Starting from a distance that induces the weak topology on measures, we define a cut distance on probability-graphons. We exhibit a tightness criterion for probability-graphons related to relative compactness in the cut distance. Finally, we prove that this topology coincides with the topology induced by the convergence in distribution of the sampled subgraphs. In the second part of this thesis, we focus on hidden Markov models indexed by trees. We show the strong consistency and asymptotic normality of the maximum likelihood estimator for these models under standard assumptions. We prove an ergodic theorem for branching Markov chains indexed by trees with general shapes. Finally, we show that for a stationary and reversible chain, the line graph is the tree shape that induces the minimal variance for the empirical mean estimator among trees with a given number of vertices
Lacour, Claire. "Estimation non paramétrique adaptative pour les chaînes de Markov et les chaînes de Markov cachées". Phd thesis, Université René Descartes - Paris V, 2007. http://tel.archives-ouvertes.fr/tel-00180107.
Texto completo da fonteDe, Almeida Rui Manuel. "Décantation dans les chaînes de Markov". Lille 1, 1986. http://www.theses.fr/1986LIL10144.
Texto completo da fonteFaure, Mathieu. "Grandes déviations autonormalisées pour des chaînes de Markov". Phd thesis, Université de Marne la Vallée, 2002. http://tel.archives-ouvertes.fr/tel-00572835.
Texto completo da fonteNoquet, Caroline. "Principe d'invariance local pour les chaînes de Markov". Lille 1, 1997. http://www.theses.fr/1997LIL10167.
Texto completo da fonteThivierge, Sylvain. "Simulation de Monte-Carlo par les chaînes de Markov". Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1999. http://www.collectionscanada.ca/obj/s4/f2/dsk1/tape10/PQDD_0004/MQ42024.pdf.
Texto completo da fonteFernandes, Clément. "Chaînes de Markov triplets et segmentation non supervisée d'images". Electronic Thesis or Diss., Institut polytechnique de Paris, 2022. http://www.theses.fr/2022IPPAS019.
Texto completo da fonteHidden Markov chains (HMC) are widely used in unsupervised Bayesian hidden discrete data restoration. They are very robust and, in spite of their simplicity, they are sufficiently efficient in many situations. In particular for image segmentation, despite their mono-dimensional nature, they are able, through a transformation of the bi-dimensional images into mono-dimensional sequences with Peano scan (PS), to give satisfying results. However, sometimes, more complex models such as hidden Markov fields (HMF) may be preferred in spite of their increased time complexity, for their better results. Moreover, hidden Markov models (the chains as well as the fields) have been extended to pairwise and triplet Markov models, which can be of interest in more complex situations. For example, when sojourn time in hidden states is not geometrical, hidden semi-Markov (HSMC) chains tend to perform better than HMC, and such is also the case for hidden evidential Markov chains (HEMC) when data are non-stationary. In this thesis, we first propose a new triplet Markov chain (TMC), which simultaneously extends HSMC and HEMC. Based on hidden triplet Markov chains (HTMC), the new hidden evidential semi-Markov chain (HESMC) model can be used in unsupervised framework, parameters being estimated with Expectation-Maximization (EM) algorithm. We validate its interest through some experiments on synthetic data. Then we address the problem of mono-dimensionality of the HMC with PS model in image segmentation by introducing the “contextual” Peano scan (CPS). It consists in associating to each index in the HMC obtained from PS, two observations on pixels which are neighbors of the pixel considered in the image, but are not its neighbors in the HMC. This gives three observations on each point of the Peano scan, which leads to a new conditional Markov chain (CMC) with a more complex structure, but whose posterior law is still Markovian. Therefore, we can apply the usual parameter estimation method: Stochastic Expectation-Maximization (SEM), as well as study unsupervised segmentation Marginal Posterior Mode (MPM) so obtained. The CMC with CPS based supervised and unsupervised MPM are compared to the classic scan based HMC-PS and the HMF through experiments on artificial images. They improve notably the former, and can even compete with the latter. Finally, we extend the CMC-CPS to Pairwise Conditional Markov (CPMC) chains and two particular triplet conditional Markov chain: evidential conditional Markov chains (CEMC) and conditional semi-Markov chains (CSMC). For each of these extensions, we show through experiments on artificial images that these models can improve notably their non conditional counterpart, as well as the CMC with CPS, and can even compete with the HMF. Beside they allow the generality of markovian triplets to better play its part in image segmentation, while avoiding the substantial time complexity of triplet Markov fields
Romaskevich, Olga. "Dynamique des systèmes physiques, formes normales et chaînes de Markov". Thesis, Lyon, 2016. http://www.theses.fr/2016LYSEN043/document.
Texto completo da fonteThis thesis deals with the questions of asymptotic behavior of dynamical systems and consists of six independent chapters. In the first part of this thesis we consider three particular dynamical systems. The first two chapters deal with the models of two physical systems: in the first chapter, we study the geometric structure and limit behavior of Arnold tongues of the equation modeling a Josephson contact; in the second chapter, we are interested in the Lagrange problem of establishing the asymptotic angular velocity of the swiveling arm on the surface. The third chapter deals with planar geometry of an elliptic billiard.The forth and fifth chapters are devoted to general methods of studying the asymptotic behavior of dynamical systems. In the forth chapter we prove the convergence of markovian spherical averages for free group actions on a probablility space. In the fifth chapter we provide a normal form for skew-product diffeomorphisms that can be useful in the study of strange attractors of dynamical systems
RAFI, Selwa. "Chaînes de Markov cachées et séparation non supervisée de sources". Phd thesis, Institut National des Télécommunications, 2012. http://tel.archives-ouvertes.fr/tel-00995414.
Texto completo da fonteCOT, CECILE. "Méthodes d'accélération pour les chaînes de Markov à transitions exponentielles". Paris 11, 1998. http://www.theses.fr/1998PA112325.
Texto completo da fonteLanchantin, Pierre. "Chaînes de Markov triplets et segmentation non supervisée de signaux". Evry, Institut national des télécommunications, 2006. http://www.theses.fr/2006TELE0012.
Texto completo da fonteThe aim of this thesis is to propose original methods of unsupervised signal and image segmentation , based on triplet Markov and partially pairwise Markov models. We first describe different models with increasing generality and develop inference and parameters estimation algorithms in the monodimensional case ( chains). Then we propose and study particular cases of triplet partially Markov chains, starting with a model of pairwise partially Markov chains to the segmentation of centured gaussian processes with long correlation noise. The segmentation of centured gaussian processes with long correlation noise. Finally, we propose a triplet Markov chains model adapted to the segmentation of non stationary hidden processes. We also study the extension possibilites of classical probabilistic models ( chains and trees) in an evidential model, where the posterior hidden process distribution is given by the Dempster-Shafer fusion and in a "fuzzy "model in which the mebership function is fuzzy
Thébaud, Olivier. "Emploi des chaînes de Markov dérivantes dans l'étude du génome". Paris 5, 2001. http://www.theses.fr/2001PA05S008.
Texto completo da fonteRafi, Selwa. "Chaînes de Markov cachées et séparation non supervisée de sources". Thesis, Evry, Institut national des télécommunications, 2012. http://www.theses.fr/2012TELE0020/document.
Texto completo da fonteThe restoration problem is usually encountered in various domains and in particular in signal and image processing. It consists in retrieving original data from a set of observed ones. For multidimensional data, the problem can be solved using different approaches depending on the data structure, the transformation system and the noise. In this work, we have first tackled the problem in the case of discrete data and noisy model. In this context, the problem is similar to a segmentation problem. We have exploited Pairwise and Triplet Markov chain models, which generalize Hidden Markov chain models. The interest of these models consist in the possibility to generalize the computation procedure of the posterior probability, allowing one to perform bayesian segmentation. We have considered these methods for two-dimensional signals and we have applied the algorithms to retrieve of old hand-written document which have been scanned and are subject to show through effect. In the second part of this work, we have considered the restoration problem as a blind source separation problem. The well-known "Independent Component Analysis" (ICA) method requires the assumption that the sources be statistically independent. In practice, this condition is not always verified. Consequently, we have studied an extension of the ICA model in the case where the sources are not necessarily independent. We have introduced a latent process which controls the dependence and/or independence of the sources. The model that we propose combines a linear instantaneous mixing model similar to the one of ICA model and a probabilistic model on the sources with hidden variables. In this context, we show how the usual independence assumption can be weakened using the technique of Iterative Conditional Estimation to a conditional independence assumption
El, Haddad Rami. "Méthodes quasi-Monte Carlo de simulation des chaînes de Markov". Chambéry, 2008. http://www.theses.fr/2008CHAMS062.
Texto completo da fonteMonte Carlo (MC) methods are probabilistic methods based on the use of random numbers in repeated simulations to estimate some parameter. Their deterministic versions are called Quasi-Monte Carlo (QMC) methods. The idea is to replace pseudo-random points by deterministic quasi-random points (also known as low-discrepancy point sets or sequences). In this work, we propose and analyze QMC-based algorithms for the simulation of multidimensional Markov chains. The quasi-random points we use are (T,S)-sequences in base B. After recalling the principles of MC and QMC methods and their main properties, we introduce some plain financial models, to serve in the following as numerical examples to test the convergence of the proposed schemes. We focus on problems where the exact solution is known, in order to be able to compute the error and to compare the efficiency of the various schemes In a first part, we consider discrete-time Markov chains with S-dimensional state spaces. We propose an iterative QMC scheme for approximating the distribution of the chain at any time. The scheme uses a (T,S+1)-sequence in base b for the transitions. Additionally, one needs to re-order the copies of the chain according to their successive components at each time-step. We study the convergence of the scheme by making some assumptions on the transition matrix. We assess the accuracy of the QMC algorithm through financial examples. The results show that the new technique is more efficient than the traditional MC approach. Then, we propose a QMC algorithm for the simulation of Markov chains with multidimensional continuous state spaces. The method uses the same re-ordering step as in the discrete setting. We provide convergence results in the case of one dimensional chains and then in the case of multidimensional chains, by making additional assumptions. We illustrate the convergence of the algorithm through numerical experiments. The results show that the new method converges faster than the MC algorithm. In the last part, we consider the problem of the diffusion equation in a spatially nonhomogeneous medium. We use a random walk algorithm, in conjunction with a correction of the Gaussian Steplength. We write a QMC variant of the algorithm, by adapting the principles seen for the simulation of the Markov chains. We test the method in dimensions 1, 2 and 3 on a problem involving the diffusion of calcium ions in a biological medium. In all the simulations, the results of QMC computations show a strong improvement over MC outcomes. Finally, we give some perspectives and directions for future work
Clémençon, Stéphan. "Méthodes d'ondelettes pour la statistique non paramétrique des chaînes de Markov". Paris 7, 2000. http://www.theses.fr/2000PA077042.
Texto completo da fonteAit, El Fquih Boujemaa. "Estimation bayésienne non supervisée dans les chaînes de Markov triplets continues". Evry, Institut national des télécommunications, 2007. http://www.theses.fr/2007TELE0014.
Texto completo da fonteA triplet Markov chain (TMC) is a stochastic dynamical model in which the state x, the observation y, and a third process r jointly form a vectorial Markov chain. This model is a generalization of the classical hidden Markov chain (HMC) model. The work of this thesis is devoted to the restoration problem and the parameter estimation problem in continuous TMC model. We propose filtering and fixed-interval Bayesian smoothing algorithms. In the particular case of Gaussian TMC, some of these algorithms extend to the TMC framework some Kalman type filtering or smoothing algorithms previously derived in the state-space framework ; however some algorithms remain original. We also propose for the general case sequential Monte Carlo based restoration algorithms. Some of these algorithms extend to TMC particle filtering or smoothing algorithms which were originally introduced in non linear and/or non Gaussian state-space systems ; some other algorithms remain original. We next adress the unsupervised case and we propose an EM parameter estimation algorithm. We finally adress a blind turbo-equalization problem in the presence of Inter-Symbol Interferences. The proposed equalizer is a fixed-lag sequential Monte Carlo smoothing algorithm
Barbu, Vlad. "Estimation des chaînes semi-markoviennes et des chaînes semi-markoviennes cachées en vue d'applications en fiabilité et en biologie". Compiègne, 2005. http://www.theses.fr/2005COMP1568.
Texto completo da fonteThe first part of my thesis concerns the discrete time semi-Markov models and the associated nonparametric estimation. The obtained results are used for deriving estimators of the systems reliability and of the associated measures. The asymptotic properties of the estimators are studied. An example illustrates how to practically compute the reliability indicators. The second part of my thesis is devoted to the estimation of hidden semi-Markov models. The asymptotic properties of the estimators are studied and an EM algorithm is proposed. An application in genetics for detecting the CpG islands in a DNA sequence shows the interest of our researches
Poly, Guillaume. "Formes de Dirichlet et applications en théorie ergodique des chaînes de Markov". Phd thesis, Université Paris-Est, 2011. http://tel.archives-ouvertes.fr/tel-00690724.
Texto completo da fonteAbbassi, Noufel. "Chaînes de Markov triplets et filtrage optimal dans les systemes à sauts". Phd thesis, Institut National des Télécommunications, 2012. http://tel.archives-ouvertes.fr/tel-00873630.
Texto completo da fonteBen, Mabrouk Mohamed. "Modèles de Markov triplets en restauration des signaux". Phd thesis, Institut National des Télécommunications, 2011. http://tel.archives-ouvertes.fr/tel-00694128.
Texto completo da fonteVergne, Nicolas. "Chaînes de Markov régulées et approximation de Poisson pour l'analyse de séquences biologiques". Phd thesis, Université d'Evry-Val d'Essonne, 2008. http://tel.archives-ouvertes.fr/tel-00322434.
Texto completo da fonteΠt/n = (1-t/n) Π0 + t/n Π1.
Cette modélisation correspond à une évolution douce entre deux états. Par exemple cela peut traduire la transition entre deux régimes d'un chaîne de Markov cachée, qui pourrait parfois sembler trop brutale. Ces modèles peuvent donc être vus comme une alternative mais aussi comme un outil complémentaire aux modèles de Markov cachés. Tout au long de ce travail, nous avons considéré des dérives polynomiales de tout degré ainsi que des dérives par splines polynomiales : le but de ces modèles étant de les rendre plus flexibles que ceux des polynômes. Nous avons estimé nos modèles de multiples manières puis évalué la qualité de ces estimateurs avant de les utiliser en vue d'applications telle la recherche de mots exceptionnels. Nous avons mis en oeuvre le software DRIMM (bientôt disponible à http://stat.genopole.cnrs.fr/sg/software/drimm/, dédié à l'estimation de nos modèles. Ce programme regroupe toutes les possibilités offertes par nos modèles, tels le calcul des matrices en chaque position, le calcul des lois stationnaires, des distributions de probabilité en chaque position... L'utilisation de ce programme pour la recherche des mots exceptionnels est proposée dans des programmes auxiliaires (disponibles sur demande).
Plusieurs perspectives à ce travail sont envisageables. Nous avons jusqu'alors décidé de faire varier la matrice seulement en fonction de la position, mais nous pourrions prendre en compte des covariables tels le degré d'hydrophobicité, le pourcentage en gc, un indicateur de la structure des protéines (hélice α, feuillets β...). Nous pourrions aussi envisager de mêler HMM et variation continue, où sur chaque région, au lieu d'ajuster un modèle de Markov, nous ajusterions un modèle de chaînes de Markov régulées.
Rouan, Lauriane. "Apports des chaînes de Markov cachées à l'analyse de données de capture-recapture". Montpellier 2, 2007. http://www.theses.fr/2007MON20188.
Texto completo da fonteParis, Sébastien. "Extraction Automatique de Pistes Fréquentielles en Sonar Passif par Chaînes de Markov Cachées". Toulon, 2000. http://www.theses.fr/2000TOUL0013.
Texto completo da fonteThe frequency vs. Time image called lofargram in any passive sonar system is the key of the downstream information processing : the operator will investigate on this representation in order to classify the tar¬gets of interest and/or to track the targets that have generated them by the so-called Target Motion Analysis (TMA). Both need to use frequency lines. For TMA, constant but Doppler-shifted frequencies are necessary. Conversely, for classification, the fluctuations of frequency help the discrimination between targets. In this thesis, we are concerned by unstable frequency line tracker (FLT). The role of such a tracker is to estimate frequency lines. The fundamental problems of FLT come from the unknown num¬ber of lines, their dates of birth and death, their respective SNR's and the case of crossing lines. We propose a method based on Markov Chain modeling : each frequency line is assumed to follow a random walk; the state space is the lofargram frequency cell set vs. A discrete and finite set of slopes. The data are composed by each lofargram line. When one unique line is present, we propose a scaled version of algorithms encountered in Hidden Markov Models (HMM) literature : Viterbi algorithm (VA), Forward (F), Forward-Backward (FB). In case of several frequency lines, we derive a new scaled FB algorithm in which each probability is conditioned by the exclusive event : two lines cannot be simultaneously in the same spot. The algorithm works in two passes : first, the lines are extracted from the beginning to the end of the lofargram; then, we estimate the dates of birth and death of each of them. When those dates are equal, the line is discarded. Therefore, the number of lines must be a priori overestimated. Trials on synthetic but also real data have been conducted and allow us to conclude that this algorithm performs very correctly (in a operational sense)
Olmi, Christophe. "Contribution à l'évaluation de la fiabilité des chaînes polyphasées de conversion électromécanique d'énergie". Thesis, Paris, ENSAM, 2019. http://www.theses.fr/2019ENAM0013/document.
Texto completo da fonteElectrical multi-phase machines exhibit intrinsic advantages (power subdivision, weak torque ripple) compared to 3-phase machines. Multi-phase machines are appreciated for marine propulsion. They own reconfiguration capabilities due to redundancy because of their high number of phases. Those capabilities are able to improve multi-phase machines reliability by using degraded modes. Presented work proposes a methodology to quantify the multi-phase system reliability. Static converter is particularly investigated as its components are a weak point in the system. Continuous virtual bases of the components are developed to prevent quantification effects. Main stressors are identified and included in the failure rates assessment of the different system components. Markov models are used to take into account the reconfiguration consequences onto the reliability function. A coupled criterion performance-reliability is introduced to characterize degraded modes into the reliability assessment. Examples of the method application from marine environment are exhibited including their topology, mission profile and control strategy, which strongly influence the stressors. A sensitivity analysis is proposed showing the input data scattering effect onto the reliability function
Royer, Alexandre. "Evaluation de performances de réseaux de communication à l'aide de chaînes de Markov hybrides". Phd thesis, Grenoble INPG, 2006. http://tel.archives-ouvertes.fr/tel-00168342.
Texto completo da fonteDridi, Noura. "Estimation aveugle de chaînes de Markov cachées simples et doubles : Application au décodage de codes graphiques". Thesis, Evry, Institut national des télécommunications, 2012. http://www.theses.fr/2012TELE0022.
Texto completo da fonteSince its birth, the technology of barcode is well investigated for automatic identification. When reading, a barcode can be degraded by a blur , caused by a bad focalisation and/ or a camera movement. The goal of this thesis is the optimisation of the receiver of 1D and 2D barcode from hidden and double Markov model and blind statistical estimation approaches. The first phase of our work consists of modelling the original image and the observed one using Hidden Markov model. Then, new algorithms for joint blur estimation and symbol detection are proposed, which take into account the non-stationarity of the hidden Markov process. Moreover, a method to select the most relevant model of the blur is proposed, based on model selection criterion. The method is also used to estimate the blur length. Finally, a new algorithm based on the double Markov chain is proposed to deal with digital communication through a long memory channel. Estimation of such channel is not possible using the classical detection algorithms based on the maximum likelihood due to the prohibitive complexity. New algorithm giving good trade off between complexity and performance is provided
Maillard, Grégory. "Chaînes à liaisons complètes et mesures de Gibbs unidimensionnelles". Rouen, 2003. http://www.theses.fr/2003ROUES015.
Texto completo da fonteWe introduce an statistical mechanical formalism for the study of discrete-time stochastic processes (chains) with which we prove: (i) General properties of extremal chains, including triviality on the tail sigma-algebra, short-range correlations, realization via infinite-volume limits and ergodicity. (ii) Two new sufficient conditions for the uniqueness of the consistent chain. (iii) Results on loss of memory and mixing properties for chains in the Dobrushin regime. We discuss the relationship between chains and one-dimensional Gibbs measures. We consider finite-alphabet systems, possibly with a grammar. We establish conditions for a chain to define a Gibbs measure and vice versa. We discuss the equivalence of uniqueness criteria for chains and fields and we establish bounds for the continuity rates of the respective systems of finite-volume conditional probabilities. We prove a (re)construction theorem for specifications starting from single-site conditioning
Nunzi, Francois. "Autour de quelques chaines de Markov combinatoires". Thesis, Sorbonne Paris Cité, 2016. http://www.theses.fr/2016USPCC270/document.
Texto completo da fonteWe consider two types of combinatoric Markov chains. We start with Juggling Markov chains, inspired from Warrington's model. We define multivariate generalizations of the existing models, for which we give stationary mesures and normalization factors with closed-form expressions. We also investigate the case where the maximum height at which the juggler may send balls tends to infinity. We then reformulate the Markov chain in terms of integer partitions, which allows us to consider the case where the juggler interacts with infinitely many balls. Our proofs are obtained through an enriched Markov chain on set partitions. We also show that one of the models has the ultrafast convergence property : the stationary mesure is reached after a finite number of steps. In the following Chapter, we consider multivariate generalizations of those models : the juggler now juggles with balls of different weights, and when a heavy ball collides with a lighter one, this light ball is bumped to a higher position, where it might collide with a lighter one, until a ball reaches the highest position. We give closed-form expressions for the stationary mesures and the normalization factors. The last Chapter is dedicated to the stochastic sandpile model, for which we give a proof for a conjecture set by Selig : the stationary mesure does not depend on the law governing sand grains additions
Guibourg, Denis. "Théorèmes de renouvellement pour des fonctionnelles additives associées à des chaînes de Markov fortement ergodiques". Phd thesis, Université Rennes 1, 2011. http://tel.archives-ouvertes.fr/tel-00583175.
Texto completo da fonteLapuyade-Lahorgue, Jérôme. "Sur diverses extensions des chaînes de Markov cachées avec application au traitement des signaux radar". Phd thesis, Institut National des Télécommunications, 2008. http://tel.archives-ouvertes.fr/tel-00473711.
Texto completo da fonteNicolas, Pierre. "Mise au point et utilisation de modèles de chaînes de Markov cachées pour l'étude des séquences d'ADN". Evry-Val d'Essonne, 2003. http://www.theses.fr/2003EVRY0017.
Texto completo da fonteConsidering the use of self-training approaches, we developed in this thesis three domains in which we applied HMM for the bacterial genome interpretation. First, a segmentation method of DNA sequences into regions of homogeneous composition enables us to identify horizontal gene transfers on the Bacillus subtilis chromosome and also others heterogeneities levels linked to biological properties of genes. Next we developed a gene prediction software and we especially focused on small genes research. Around 30 genes smaller than 50 amino acids have been added to about 20 small genes previously biologically identified on B. Subtilis. Then we proposed a MCMC algorithm for Bayesian model selection in the context of RNA polymerase binding sites modeling
Garivier, Aurélien. "Modèles contextuels et alphabets infinis en théorie de l'information". Paris 11, 2006. http://www.theses.fr/2006PA112192.
Texto completo da fonteThis thesis explores some contemporary aspects of information theory, from source coding to issues of model selection. We first consider the problem of coding memoryless sources on a countable, infinite alphabet. As it is impossible to provide a solution which is both efficient and general, two approaches are considered: we first establish conditions under which the entropic rate can be reached, and we consider restricted classes for which tail probabilities are controlled. The second approach does not set any condition on the sources but provides a partial solution by coding only a part of the information - the pattern - which captures the repetitions in the message. In order to study more complex processes, we come back to the case of finite memory sources on a finite alphabet : it has given rise to many works and efficient algorithms like the Context Tree Weighting (CTW) Method. We show here that this method is also efficient on anon-parametric class of infinite memory sources: the renewal processes. We show then that the ideas on which CTW is based lead to a consistent estimator of the memory structure of a process, when this structure is finite. In fact, we complete the study of the BIC context tree estimator for Variable Length Markov Chains. In the last part, it is shown how similar ideas can be generalized for more complex sources on a (countable or not) infinite alphabet. We obtain consistent estimators for the order of hidden Markov models with Poisson and Gaussian emission
Prabhu, Balakrishna J. "Chaînes de Markov et processus de décision markoviens pour le contrôle de congestion et de puissance". Phd thesis, Université de Nice Sophia-Antipolis, 2005. http://tel.archives-ouvertes.fr/tel-00328111.
Texto completo da fonteGbedo, Yémalin Gabin. "Les techniques Monte Carlo par chaînes de Markov appliquées à la détermination des distributions de partons". Thesis, Université Grenoble Alpes (ComUE), 2017. http://www.theses.fr/2017GREAY059/document.
Texto completo da fonteWe have developed a new approach to determine parton distribution functions and quantify their experimental uncertainties, based on Markov Chain Monte Carlo methods.The main interest devoted to such a study is that we can replace the standard χ 2 MINUIT minimization by procedures grounded on Statistical Methods, and on Bayesian inference in particular, thus offering additional insight into the rich field of PDFs determination.After reviewing these Markov chain Monte Carlo techniques, we introduce the algorithm we have chosen to implement – namely Hybrid (or Hamiltonian) Monte Carlo. This algorithm, initially developed for lattice quantum chromodynamique, turns out to be very interesting when applied to parton distribution functions determination by global analyses ; we have shown that it allows to circumvent the technical difficulties due to the high dimensionality of the problem, in particular concerning the acceptance rate. The feasibility study performed and presented in this thesis, indicates that Markov chain Monte Carlo method can successfully be applied to the extraction of PDFs and of their experimental uncertainties
Vandekerkhove, Pierre. "Identification de l'ordre des processus ARMA stables : contribution à l'étude statistique des chaînes de Markov cachées". Montpellier 2, 1997. http://www.theses.fr/1997MON20115.
Texto completo da fonteRichard, Hugues. "Prédiction de la localisation cellulaire des protéines à l'aide de leurs séquences biologiques". Phd thesis, Université d'Evry-Val d'Essonne, 2005. http://tel.archives-ouvertes.fr/tel-00011707.
Texto completo da fonteAinsi la majorité de ce travail de thèse s'intéresse au problème de la prédiction du compartiment cellulaire d'une protéine à partir de sa séquence primaire.
Nous nous sommes attachés à proposer des alternatives descriptives aux méthodes existantes de prédiction de la localisation cellulaire en utilisant : (1) de nouveaux descripteurs issus de la séquence nucléique, (2) une approche par chaînes de Markov cachées (CMC) et arbres de décision. L'approche par CMC est justifiée biologiquement a posteriori car elle permet la modélisation de signaux d'adressage conjointement à la prise en compte de la composition globale. En outre, l'étape de classification hiérarchique par arbre améliore nettement les résultats de classification. Les résultats obtenues lors des comparaisons avec les méthodes existantes et utilisant des descripteurs fondés sur la composition globale possèdent des performances similaires.
Boudaren, Mohamed El Yazid. "Modèles graphiques évidentiels". Phd thesis, Institut National des Télécommunications, 2014. http://tel.archives-ouvertes.fr/tel-01004504.
Texto completo da fonteAit, Salaht Farah. "Chaînes de Markov Incomplètement spécifiées : analyse par comparaison stochastique et application à l'évaluation de performance des réseaux". Thesis, Versailles-St Quentin en Yvelines, 2014. http://www.theses.fr/2014VERS0018.
Texto completo da fonteThis thesis is devoted to the uncertainty in probabilistic models, how it impacts their analysis and how to apply these methods to performance analysis and network dimensioning. We consider two aspects of the uncertainty. The first consists to study a partially specified Markov chains. The missing of some transitions in the exact system because of its complexity can be solved by constructing bounding systems where worst-case transitions are defined to obtain an upper or a lower bound on the performance measures. We propose to develop new algorithms which give element-wise bounds of the steady-state distribution for the partially specified Markov chain. These algorithms are faster than the existing ones and allow us to compute element-wise bounds at each iteration.The second aspect studied concerns the problem of the measurements of real traffic trace in networks. Exact analysis of queueing networks under real traffic becomes quickly intractable due to the state explosion. Assuming the stationarity of flows, we propose to apply the stochastic comparison method to derive performance measure bounds under histogram-based traffics. We apply an algorithm based on dynamic programming to derive optimal bounding traffic histograms on reduced state spaces. Using the stochastic bound histograms and the monotonicity of the networking elements, we show how we can obtain, in a very efficient manner, guarantees on performance measures. We indeed obtain stochastic upper and lower bounds on buffer occupancy, losses, etc. The interest and the impact of our method are shown on various applications: elements of networks, AQM, queueing networks and queue with non-stationary arrival process
Muraro, Anthony. "Processus de Hawkes en temps discret avec inhibition". Electronic Thesis or Diss., Université de Toulouse (2023-....), 2024. http://www.theses.fr/2024TLSES103.
Texto completo da fonteThis thesis focuses on Hawkes processes, which are continuous-time stochastic processes whose intensity is random and depends on the entire history of the process. These processes were introduced by Hawkes (1971) to model self-exciting dynamics. A generalization of these processes involves incorporating a self-inhibition effect, for which the literature is more limited and notably lacks a necessary and sufficient criterion for the existence of a stationary version that truly accounts for the inhibitory effect of the model. In the first chapter, we conduct a numerical analysis on the simulation of the solution to a nonlinear renewal equation, which is closely related to the expectation of the conditional intensity of a Hawkes process. Additionally, we present examples of simulations of these processes in settings that do not meet the previously mentioned criterion, thereby illustrating the restrictive nature of this criterion. Next, we focus more specifically on a discrete version of Hawkes processes with inhibition. This model takes the form of an autoregressive Poisson process, where the parameter is random and depends on the past realizations of the process. We allow these parameters to take negative values to model an inhibitory effect. The second chapter examines the case where the memory of this process is of size 2. In this context, we classify the asymptotic behavior of this process for the whole range of parameters, except for boundary cases. To achieve these results, we use the formalism of Markov chains as described by Douc, Moulines and Priouret (2018), and employ Foster's (1953) criterion-type theorems via Lyapunov functions. We also take advantage of comparing the behavior of this process with the behavior of solutions to a naturally associated linear recurrence sequence. In the third chapter, we extend our study to the case where the memory of the process is of size 3, and provide some results in the general case of arbitrarily large memory. We complement our theoretical study with numerical simulations to support our conjectures and provide intuition about the long-term behavior of this process. The final chapter is ongoing work on the critical cases of the studied process. We notably use diffusion approximation tools from the book by Ethier and Kurtz (1986). We show that after an appropriate time and space rescaling, the process converges to a diffusion governed by a stochastic differential equation, which we explicitly describe
Bercu, Sophie. "Modélisation stochastique du signal écrit par chaînes de Markov cachées : application à la reconnaissance automatique de l'écriture manuscrite". Rennes 1, 1994. http://www.theses.fr/1994REN10115.
Texto completo da fonteSuparman, Suparman. "Problèmes de choix de modèles par simulation de type Monte Carlo par chaînes de Markov à sauts réversibles". Toulouse 3, 2003. http://www.theses.fr/2003TOU30005.
Texto completo da fonteBrunel, Nicolas. "Sur quelques extensions des chaînes de Markov cachées et couples : application à la segmentation non supervisée de signaux radar". Paris 6, 2005. https://tel.archives-ouvertes.fr/tel-00011302.
Texto completo da fonteAdam, Etienne. "Persistance et vitesse d'extinction pour des modèles de populations stochastiques multitypes en temps discret". Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLX019/document.
Texto completo da fonteThis thesis is devoted to the mathematical study of stochastic modelds of structured populations dynamics.In the first chapter, we introduce a discrete time stochastic process taking into account various ecological interactions between individuals, such as competition, migration, mutation, or predation. We first prove a ``law of large numbers'': where we show that if the initial population tends to infinity, then, on any finite interval of time, the stochastic process converges in probability to an underlying deterministic process. We also quantify the discrepancy between these two processes by a kind of ``central limit theorem''. Finally, we give a criterion of persistence/extinction in order to determine the long time behavior of the process. This criterion highlights a critical case which will be studied in more detail in the following chapters.In the second chapter, we give a criterion for the possible unlimited growth in the critical case mentioned above. We apply this criterion to the example of a source-sink metapopulation with two patches of type source, textit{i.e.} the population of each patch goes to extinction if we do not take into account the migration. We prove that there is a possible survival of the metapopulation.In the third chapter, we focus on the behavior of our critical process when it tends to infinity. We prove a convergence in distribution of the scaled process to a gamma distribution, and in a more general framework, by also rescaling time, we obtain a distribution limit of a function of our process to the solution of a stochastic differential equation called a squared Bessel process.In the fourth and last chapter, we study hitting times of some compact sets when our process does not tend to infinity. We give nearly optimal bounds for the tail of these hitting times. If the process goes to extinction almost surely, we deduce from these bounds precise estimates of the tail of the extinction time. Moreover, if the process is a Markov chain, we give a criterion of null recurrence or positive recurrence and in the latter case, we obtain a subgeometric convergence of its transition kernel to its invariant probability measure
Lièvre, Agnès. "Mortalité aux grands âges et espérance de vie en santé mesurée à partir des enquêtes transverso-longitudinales". Paris 7, 2004. http://www.theses.fr/2004PA077118.
Texto completo da fonteDerouault, Anne-Marie. "Modélisation d'une langue naturelle pour la désambiguation des chaînes phonétiques". Paris 7, 1985. http://www.theses.fr/1985PA077028.
Texto completo da fonteVarloot, Rémi. "Dynamic network formation". Thesis, Paris Sciences et Lettres (ComUE), 2018. http://www.theses.fr/2018PSLEE048/document.
Texto completo da fonteThis thesis focuses on the rapid mixing of graph-related Markov chains. The main contribution concerns graphs with local edge dynamics, in which the topology of a graph evolves as edges slide along one another. We propose a classification of existing models of dynamic graphs, and illustrate how evolving along a changing structure improves the convergence rate. This is complemented by a proof of the rapid mixing time for one such dynamic. As part of this proof, we introduce the partial expansion of a graph. This notion allows us to track the progression of the dynamic, from a state with poor expansion to good expansion at equilibrium. The end of the thesis proposes an improvement of the Propp and Wilson perfect sampling technique. We introduce oracle sampling, a method inspired by importance sampling that reduces the overall complexity of the Propp and Wilson algorithm. We provide a proof of correctness, and study the performance of this method when sampling independent sets from certain graphs
Brunel, Nicolas. "Sur quelques extensions des chaînes de Markov cachées et couples. Applications à la segmentation non-supervisée de signaux radar". Phd thesis, Université Pierre et Marie Curie - Paris VI, 2005. http://tel.archives-ouvertes.fr/tel-00011302.
Texto completo da fonteMuri, Florence. "Comparaison d'algorithmes d'identification de chaînes de Markov cachées et application a la détection de régions homogènes dans les séquences d'ADN". Paris 5, 1997. http://www.theses.fr/1997PA05S008.
Texto completo da fonteBoudaren, Mohamed El Yazid. "Modèles graphiques évidentiels". Electronic Thesis or Diss., Evry, Institut national des télécommunications, 2014. http://www.theses.fr/2014TELE0001.
Texto completo da fonteHidden Markov chains (HMCs) based approaches have been shown to be efficient to resolve a wide range of inverse problems occurring in image and signal processing. In particular, unsupervised segmentation of data is one of these problems where HMCs have been extensively applied. According to such models, the observed data are considered as a noised version of the requested segmentation that can be modeled through a finite Markov chain. Then, Bayesian techniques such as MPM can be applied to estimate this segmentation even in unsupervised way thanks to some algorithms that make it possible to estimate the model parameters from the only observed data. HMCs have then been generalized to pairwise Markov chains (PMCs) and triplet Markov chains (TMCs), which offer more modeling possibilities while showing comparable computational complexities, and thus, allow to consider some challenging situations that the conventional HMCs cannot support. An interesting link has also been established between the Dempster-Shafer theory of evidence and TMCs, which give to these latter the ability to handle multisensor data. Hence, in this thesis, we deal with three challenging difficulties that conventional HMCs cannot handle: nonstationarity of the a priori and/or noise distributions, noise correlation, multisensor information fusion. For this purpose, we propose some original models in accordance with the rich theory of TMCs. First, we introduce the M-stationary noise- HMC (also called jumping noise- HMC) that takes into account the nonstationary aspect of the noise distributions in an analogous manner with the switching-HMCs. Afterward, ML-stationary HMC consider nonstationarity of both the a priori and/or noise distributions. Second, we tackle the problem of non-stationary PMCs in two ways. In the Bayesian context, we define the M-stationary PMC and the MM-stationary PMC (also called switching PMCs) that partition the data into M stationary segments. In the evidential context, we propose the evidential PMC in which the realization of the hidden process is modeled through a mass function. Finally, we introduce the multisensor nonstationary HMCs in which the Dempster-Shafer fusion has been used on one hand, to model the data nonstationarity (as done in the hidden evidential Markov chains) and on the other hand, to fuse the information provided by the different sensors (as in the multisensor hidden Markov fields context). For each of the proposed models, we describe the associated segmentation and parameters estimation procedures. The interest of each model is also assessed, with respect to the former ones, through experiments conducted on synthetic and real data
Lefaucheux, Engel. "Controlling information in probalistic systems". Thesis, Rennes 1, 2018. http://www.theses.fr/2018REN1S057/document.
Texto completo da fonteThe control of the information given by a system has seen increasing importance recently with the multiplication of communicating systems. This control can be used in order to disclose an information of the system, or, oppositely, to hide one. Diagnosis for instance tries to determine from the observation produced by the system whether a fault occurred within it or not. In this PhD, we establish formal foundations to the analysis of the diagnosis problems for stochastic models. We then study these problems in multiple framework (finite/infinite, passive/active)