Dans cette thèse, nous nous intéressons à la représentation et à la prise en compte des incertitudes dans les systèmes de prévision hydrologique probabilistes à moyen-terme. Ces incertitudes proviennent principalement de deux sources : (1) de l’imperfection des prévisions météorologiques (utilisées en intrant de ces systèmes) et (2) de l’imperfection de la représentation du processus hydrologique par le simulateur pluie-débit (SPQ) (au coeur de ces systèmes).La performance d’un système de prévision probabiliste s’évalue par la précision de ses prévisions conditionnellement à sa fiabilité. L’approche statistique que nous suivons procure une garantie de fiabilité à condition que les hypothèses qu’elle implique soient réalistes. Nous cherchons de plus à gagner en précision en incorporant des informations auxiliaires.Nous proposons, pour chacune des sources d’incertitudes, une méthode permettant cette incorporation : (1) un post-traitement des prévisions météorologiques s’appuyant sur la propriété statistique d’échangeabilité et permettant la prise en compte de plusieurs sources de prévisions, ensemblistes ou déterministes ; (2) un post-traitement hydrologique utilisant les variables d’état des SPQ par le biais d’un modèle Probit arbitrant entre deux régimes hydrologiques interprétables et permettant ainsi de représenter une incertitude à variance hétérogène.Ces deux méthodes montrent de bonnes capacités d’adaptation aux cas d’application variés fournis par EDF et Hydro-Québec, partenaires et financeurs du projet. Elles présentent de plus un gain en simplicité et en formalisme par rapport aux méthodes opérationnelles tout en montrant des performances similaires. / In this thesis, we are interested in representing and taking into account uncertainties in medium term probabilistic hydrological prediction systems.These uncertainties mainly come from two sources: (1) from the imperfection of meteorological forecasts (used as inputs to these systems) and (2) from the imperfection of the representation of the hydrological process by the rainfall-runoff simulator (RRS) (at the heart of these systems).The performance of a probabilistic forecasting system is assessed by the sharpness of its predictions conditional on its reliability. The statistical approach we follow provides a guarantee of reliability if the assumptions it implies are complied with. We are also seeking to incorporate auxilary information to get sharper.We propose, for each source of uncertainty, a method enabling this incorporation: (1) a meteorological post-processor based on the statistical property of exchangeability and enabling to take into account several (ensemble or determistic) forecasts; (2) a hydrological post-processor using the RRS state variables through a Probit model arbitrating between two interpretable hydrological regimes and thus representing an uncertainty with heterogeneous variance.These two methods demonstrate adaptability on the various application cases provided by EDF and Hydro-Québec, which are partners and funders of the project. Those methods are moreover simpler and more formal than the operational methods while demonstrating similar performances.

L’algorithme EM (Dempster et al., 1977) permet de construire une séquence d’estimateurs qui converge vers l’estimateur de vraisemblance maximale pour des modèles à données manquantes pour lesquels l’estimateur du maximum de vraisemblance n’est pas calculable. Cet algorithme est remarquable compte tenu de ses nombreuses applications en apprentissage statistique. Toutefois, il peut avoir un lourd coût computationnel. Les auteurs Cappé et Moulines (2009) ont proposé une version en ligne de cet algorithme pour les modèles appartenant à la famille exponentielle qui permet de faire des gains d’efficacité computationnelle importants en présence de grands jeux de données. Cependant, le calcul de l’espérance a posteriori de la statistique exhaustive, qui est nécessaire dans la version de Cappé et Moulines (2009), est rarement possible pour des modèles complexes et/ou lorsque la dimension des données manquantes est grande. On doit alors la remplacer par un estimateur. Plusieurs questions se présentent naturellement : les résultats de convergence de l’algorithme initial restent-ils valides lorsqu’on remplace l’espérance par un estimateur ? En particulier, que dire de la normalité asymptotique de la séquence des estimateurs ainsi créés, de la variance asymptotique et de la vitesse de convergence ? Comment la variance de l’estimateur de l’espérance se reflète-t-elle sur la variance asymptotique de l’estimateur EM? Peut-on travailler avec des estimateurs de type Monte-Carlo ou MCMC? Peut-on emprunter des outils populaires de réduction de variance comme les variables de contrôle ? Ces questions seront étudiées à l’aide d’exemples de modèles à variables latentes. Les contributions principales de ce mémoire sont une présentation unifiée des algorithmes EM d’approximation stochastique, une illustration de l’impact au niveau de la variance lorsque l’espérance a posteriori est estimée dans les algorithmes EM en ligne et l’introduction d’algorithmes EM en ligne permettant de réduire la variance supplémentaire occasionnée par l’estimation de l’espérance a posteriori. / The EM algorithm Dempster et al. (1977) yields a sequence of estimators that converges to the maximum likelihood estimator for missing data models whose maximum likelihood estimator is not directly tractable. The EM algorithm is remarkable given its numerous applications in statistical learning. However, it may suffer from its computational cost. Cappé and Moulines (2009) proposed an online version of the algorithm in models whose likelihood belongs to the exponential family that provides an upgrade in computational efficiency in large data sets. However, the conditional expected value of the sufficient statistic is often intractable for complex models and/or when the missing data is of a high dimension. In those cases, it is replaced by an estimator. Many questions then arise naturally: do the convergence results pertaining to the initial estimator hold when the expected value is substituted by an estimator? In particular, does the asymptotic normality property remain in this case? How does the variance of the estimator of the expected value affect the asymptotic variance of the EM estimator? Are Monte-Carlo and MCMC estimators suitable in this situation? Could variance reduction tools such as control variates provide variance relief? These questions will be tackled by the means of examples containing latent data models. This master’s thesis’ main contributions are the presentation of a unified framework for stochastic approximation EM algorithms, an illustration of the impact that the estimation of the conditional expected value has on the variance and the introduction of online EM algorithms which reduce the additional variance stemming from the estimation of the conditional expected value.

Diplomová práce se zabývá zobecněným exponenciálním rozdělením jako alternativou k Weibullovu a log-normálnímu rozdělení. Jsou popsány základní charakteristiky tohoto rozdělení a metody odhadu parametrů. Samostatná kapitola je věnována testům dobré shody. Druhá část práce se zabývá cenzorovanými výběry. Jsou uvedeny ukázkové příklady pro exponenciální rozdělení. Dále je studován případ cenzorování typu I zleva, který dosud nebyl publikován. Pro tento speciální případ jsou provedeny simulace s podrobným popisem vlastností a chování. Dále je pro toto rozdělení odvozen EM algoritmus a jeho efektivita je porovnána s metodou maximální věrohodnosti. Vypracovaná teorie je aplikována pro analýzu environmentálních dat.

[pt] A reserva IBNR é uma reserva de suma importância para as seguradoras. Seu cálculo tem sido realizado por métodos, em sua grande maioria, determinísticos, tradicionalmente aplicados a informações de sinistros agrupadas num formato particular intitulado triangulo de run-off. Esta forma de cálculo foi muito usada por décadas por sua simplicidade e pela limitação da capacidade de processamento computacional existente. Hoje, com o grande avanço dessa capacidade, não haveria necessidade de deixar de investigar informações relevantes que podem ser perdidas com agrupamento dos dados. Muitas são as deficiências dos métodos tradicionais apontadas na literatura e o uso de informação detalhada tem sido apontado por alguns artigos como a fonte para superação dessas deficiências. Outra busca constante nas metodologias propostas para cálculo da IBNR é pela obtenção de boas medidas de precisão das estimativas obtidas por eles. Neste ponto, sobre o uso de dados detalhados, há a expectativa de obtenção de medidas de precisão mais justas, já que se tem mais dados. Inspirada em alguns artigos já divulgados com propostas para modelagem desses dados não agrupados esta dissertação propõe um novo modelo, avaliando sua capacidade de predição e ganho de conhecimento a respeito do processo de ocorrência e aviso de sinistros frente ao que se pode obter a partir dos métodos tradicionais aplicados à dados de quantidade para obtenção da quantidade de sinistros IBNR e sua distribuição. / [en] The IBNR reserve is a reserve of paramount importance for insurers. Its calculation has been accomplished by methods, mostly, deterministic, traditionally applied to claims grouped information in a particular format called run-off triangle . This method of calculation was very adequate for decades because of its simplicity and the limited computational processing capacity existing in the past. Today, with the breakthrough of this capacity, no waiver to investigating relevant information that may be lost with grouping data would be need. Many flaws of the traditional methods has been mentioned in the literature and the use of detailed information has been pointed as a form of overcoming these deficiencies. Another frequent aim in methodologies proposed for the calculation of IBNR is get a good measure of the accuracy of the estimates obtained by them and that is another expectation about the use of detailed data, since if you got more data you could get better measures. Inspired by some articles already published with proposals for modeling such not grouped data, this dissertation proposes a new model and evaluate its predictive ability and gain of knowledge about the process of occurrence and notice of the claim against that one can get from the traditional methods applied to data of amount of claims for obtain the amount of IBNR claims and their distribution.

No description available.

No description available.

<p>In this thesis, inferential methods for one-shot device testing data from accelerated life-test are developed. Due to constraints on time and budget, accelerated life-tests are commonly used to induce more failures within a reasonable amount of test-time for obtaining more lifetime information that will be especially useful in reliability analysis. One-shot devices, which can be used only once as they get destroyed immediately after testing, yield observations only on their condition and not on their real lifetimes. So, only binary response data are observed from an one-shot device testing experiment. Since no failure times of units are observed, we use the EM algorithm for determining the maximum likelihood estimates of the model parameters. Also, inference for the reliability at a mission time and the mean lifetime at normal operating conditions are also developed.</p> <p>The thesis proceeds as follows. Chapter 2 considers the exponential distribution with single-stress relationship and develops inferential methods for the model parameters, the reliability and the mean lifetime. The results obtained by the EM algorithm are compared with those obtained from the Bayesian approach. A one-shot device testing data is analyzed by the proposed method and presented as an illustrative example. Next, in Chapter 3, the exponential distribution with multiple-stress relationship is considered and corresponding inferential results are developed. Jackknife technique is described for the bias reduction in the developed estimates. Interval estimation for the reliability and the mean lifetime are also discussed based on observed information matrix, jackknife technique, parametric bootstrap method, and transformation technique. Again, we present an example to illustrate all the inferential methods developed in this chapter. Chapter 4 considers the point and interval estimation for the one-shot device testing data under the Weibull distribution with multiple-stress relationship and illustrates the application of the proposed methods in a study involving the development of tumors in mice with respect to risk factors such as sex, strain of offspring, and dose effects of benzidine dihydrochloride. A Monte Carlo simulation study is also carried out to evaluate the performance of the EM estimates for different levels of reliability and different sample sizes. Chapter 5 describes a general algorithm for the determination of the optimal design of an accelerated life-test plan for one-shot device testing experiment. It is based on the asymptotic variance of the estimated reliability at a specific mission time. A numerical example is presented to illustrate the application of the algorithm. Finally, Chapter 6 presents some concluding remarks and some additional research problems that would be of interest for further study.</p> / Doctor of Philosophy (PhD)

<p>Left truncation arises because in many situations, failure of a unit is observed only if it fails after a certain period. In many situations, the units under study may not be followed until all of them fail and the experimenter may have to stop at a certain time when some of the units may still be working. This introduces right censoring into the data. Some commonly used lifetime distributions are lognormal, Weibull and gamma, all of which are special cases of the flexible generalized gamma family. Likelihood inference via the Expectation Maximization (EM) algorithm is used to estimate the model parameters of lognormal, Weibull, gamma and generalized gamma distributions, based on left truncated and right censored data. The asymptotic variance-covariance matrices of the maximum likelihood estimates (MLEs) are derived using the missing information principle. By using the asymptotic variances and the asymptotic normality of the MLEs, asymptotic confidence intervals for the parameters are constructed. For comparison purpose, Newton-Raphson (NR) method is also used for the parameter estimation, and asymptotic confidence intervals corresponding to the NR method and parametric bootstrap are also obtained. Through Monte Carlo simulations, the performance of all these methods of inference are studied. With regard to prediction analysis, the probability that a right censored unit will be working until a future year is estimated, and an asymptotic confidence interval for the probability is then derived by the delta-method. All the methods of inference developed here are illustrated with some numerical examples.</p> / Doctor of Philosophy (PhD)

<p>Recently, the Conway-Maxwell Poisson (COM-Poisson) cure rate model has been proposed which includes as special cases some of the well-known cure rate models discussed in the literature. Data obtained from cancer clinical trials are often right censored and the expectation maximization (EM) algorithm can be efficiently used for the determination of the maximum likelihood estimates (MLEs) of the model parameters based on right censored data.</p> <p>By assuming the lifetime distribution to be exponential, lognormal, Weibull, and gamma, the necessary steps of the EM algorithm are developed for the COM-Poisson cure rate model and some of its special cases. The inferential method is examined by means of an extensive simulation study. Model discrimination within the COM-Poisson family is carried out by likelihood ratio test as well as by information-based criteria. Finally, the proposed method is illustrated with a cutaneous melanoma data on cancer recurrence. As the lifetime distributions considered are not nested, it is not possible to carry out a formal statistical test to determine which among these provides an adequate fit to the data. For this reason, the wider class of generalized gamma distributions is considered which contains all of the above mentioned lifetime distributions as special cases. The steps of the EM algorithm are then developed for this general class of distributions and a simulation study is carried out to evaluate the performance of the proposed estimation method. Model discrimination within the generalized gamma family is carried out by likelihood ratio test and information-based criteria. Finally, for the considered cutaneous melanoma data, the two-way flexibility of the COM-Poisson family and the generalized gamma family is utilized to carry out a two-way model discrimination to select a parsimonious competing cause distribution along with a suitable choice of a lifetime distribution that provides the best fit to the data.</p> / Doctor of Philosophy (PhD)

Researchers conduct life testing on objects of interest in an attempt to determine their life distribution as a means of studying their reliability (or survivability). Determining the life distribution of the objects under study helps manufacturers to identify potential faults, and to improve quality. Researchers sometimes conduct accelerated life tests (ALTs) to ensure that failure among the tested units is earlier than what could result under normal operating (or environmental) conditions. Moreover, such experiments allow the experimenters to examine the effects of high levels of one or more stress factors on the lifetimes of experimental units. Examples of stress factors include, but not limited to, cycling rate, dosage, humidity, load, pressure, temperature, vibration, voltage, etc. A special class of ALT is step-stress accelerated life testing. In this type of experiments, the study sample is tested at initial stresses for a given period of time. Afterwards, the levels of the stress factors are increased in agreement with prefixed points of time called stress-change times. In practice, time and resources are limited; thus, any experiment is expected to be constrained to a deadline which is called a termination time. Hence, the observed information may be subjected to Type-I censoring. This study discusses maximum likelihood inferential methods for the parameters of multiple step-stress models from a generalized Birnbaum-Saunders distribution under time constraint alongside other inference-related problems. A couple of general inference frameworks are studied; namely, the observed likelihood (OL) framework, and the expectation-maximization (EM) framework. The last-mentioned framework is considered since there is a possibility that Type-I censored data are obtained. In the first framework, the scoring algorithm is used to get the maximum likelihood estimators (MLEs) for the model parameters. In the second framework, EM-based algorithms are utilized to determine the required MLEs. Obtaining observed information matrices under both frameworks is also discussed. Accordingly, asymptotic and bootstrap-based interval estimators for the model parameters are derived. Model discrimination within the considered generalized Birnbaum-Saunders distribution is carried out by likelihood ratio test as well as by information-based criteria. The discussed step-stress models are illustrated by analyzing three real-life datasets. Accordingly, establishing optimal multiple step-stress test plans based on cost considerations and three optimality criteria is discussed. Since maximum likelihood estimators are obtained by numerical optimization that involves maximizing some objective functions, optimization methods used, and their software implementations in R are discussed. Because of the computational aspects are in focus in this study, the benefits of parallel computing in R, as a high-performance computational approach, are briefly addressed. Numerical examples and Monte Carlo simulations are used to illustrate and to evaluate the methods presented in this thesis. / Thesis / Doctor of Science (PhD)