No description available.

No description available.

A modelagem da demanda de água residencial fornece importantes subsídios ao dimensionamento e gerenciamento de redes de abastecimento de água. O comportamento desta demanda pode ser descrito através de processos estocásticos, caracterizados pela ocorrência de pulsos retangulares de demanda de água ao longo do tempo. Nesse contexto, este trabalho de pesquisa teve como objetivos monitorar e modelar a demanda de água residencial através dos modelos estocásticos dos pulsos retangulares de Neyman Scott (NSRP) e dos Pulsos Totais (OP). Enquanto que no modelo NSRP há a tentativa de simular a demanda de água através da representação dos seus constituintes elementares, no modelo OP busca-se a representação direta da demanda de água agregada dos usuários finais, como observada nos hidrômetros. A calibração e a validação dos modelos foram feitas a partir do monitoramento do consumo de água de quatro residências localizadas na cidade de São Carlos, caracterizadas por perfis de abastecimento distintos. Para tanto, foram desenvolvidos dispositivos dataloggers, que associados aos sensores/emissores de pulsos dos hidrômetros, permitiram o monitoramento do consumo de água ao longo do tempo dos usuários residenciais individuais. Durante a elaboração da pesquisa, foram observados efeitos negativos nas modelagens relacionados à influência dos reservatórios domiciliares (caixas d\'água) no perfil temporal do consumo de água das residências. Buscando mitigar esses efeitos, foram propostas modificações nas etapas de calibração e de geração das séries sintéticas de demanda de água. De uma forma geral, observou-se que as modificações propostas contribuíram para que as séries sintéticas geradas a partir dos modelos NSRP e OP reproduzissem de forma mais acurada as estatísticas das séries observadas, principalmente com relação às intensidades e durações das demandas simuladas. Apesar de as versões modificadas dos modelos NSRP e OP apresentarem desempenho similar na reprodução das médias, variâncias e covariâncias das séries observadas, o modelo OP reproduziu de forma mais consistente os volumes consumidos diários observados. / The modelling of residential water demand provides important subsidies for the design and management of water supply networks. The behavior of this demand can be described through stochastic processes, characterized by the occurrence of rectangular pulses of water demand over time. In this context, the objectives of this research were to monitor and model residential water demand using the Neyman Scott Rectangular Pulse model (NSRP) and Overall Pulse model (OP). While in the NSRP model there is the attempt to simulate the water demand through the representation of its elementary constituents, the OP model aims to direct represent the aggregate water demand of the end users, as observed in water meters. The calibration and validation of the models were done by monitoring the water consumption of four residences located in the city of São Carlos, characterized by different supply profiles. To this end, dataloggers were developed, which, coupled with sensors/pulse emitters and water meters, allowed the monitoring of water consumption over time of individual residential users. During the research, negative effects were observed in the models, related to the influence of the domestic reservoirs on the temporal patter of water consumption of the residences. To mitigate these effects, modifications were proposed in the calibration and generation stages of the synthetic water demand generation series. In general, it was observed that these proposed modifications contributed to a more accurately reproduction of the observed series statistics by the OP and NSRP synthetic series, especially regarding the intensities and durations of the simulated demands. Although the modified versions of the NSRP and OP models presented similar performance in the reproduction of the means, variances and covariance of the observed series, the OP model reproduced in a more consistent way the observed daily consumed volumes.

A modelagem da demanda de água residencial fornece importantes subsídios ao dimensionamento e gerenciamento de redes de abastecimento de água. O comportamento desta demanda pode ser descrito através de processos estocásticos, caracterizados pela ocorrência de pulsos retangulares de demanda de água ao longo do tempo. Nesse contexto, este trabalho de pesquisa teve como objetivos monitorar e modelar a demanda de água residencial através dos modelos estocásticos dos pulsos retangulares de Neyman Scott (NSRP) e dos Pulsos Totais (OP). Enquanto que no modelo NSRP há a tentativa de simular a demanda de água através da representação dos seus constituintes elementares, no modelo OP busca-se a representação direta da demanda de água agregada dos usuários finais, como observada nos hidrômetros. A calibração e a validação dos modelos foram feitas a partir do monitoramento do consumo de água de quatro residências localizadas na cidade de São Carlos, caracterizadas por perfis de abastecimento distintos. Para tanto, foram desenvolvidos dispositivos dataloggers, que associados aos sensores/emissores de pulsos dos hidrômetros, permitiram o monitoramento do consumo de água ao longo do tempo dos usuários residenciais individuais. Durante a elaboração da pesquisa, foram observados efeitos negativos nas modelagens relacionados à influência dos reservatórios domiciliares (caixas d\'água) no perfil temporal do consumo de água das residências. Buscando mitigar esses efeitos, foram propostas modificações nas etapas de calibração e de geração das séries sintéticas de demanda de água. De uma forma geral, observou-se que as modificações propostas contribuíram para que as séries sintéticas geradas a partir dos modelos NSRP e OP reproduzissem de forma mais acurada as estatísticas das séries observadas, principalmente com relação às intensidades e durações das demandas simuladas. Apesar de as versões modificadas dos modelos NSRP e OP apresentarem desempenho similar na reprodução das médias, variâncias e covariâncias das séries observadas, o modelo OP reproduziu de forma mais consistente os volumes consumidos diários observados. / The modelling of residential water demand provides important subsidies for the design and management of water supply networks. The behavior of this demand can be described through stochastic processes, characterized by the occurrence of rectangular pulses of water demand over time. In this context, the objectives of this research were to monitor and model residential water demand using the Neyman Scott Rectangular Pulse model (NSRP) and Overall Pulse model (OP). While in the NSRP model there is the attempt to simulate the water demand through the representation of its elementary constituents, the OP model aims to direct represent the aggregate water demand of the end users, as observed in water meters. The calibration and validation of the models were done by monitoring the water consumption of four residences located in the city of São Carlos, characterized by different supply profiles. To this end, dataloggers were developed, which, coupled with sensors/pulse emitters and water meters, allowed the monitoring of water consumption over time of individual residential users. During the research, negative effects were observed in the models, related to the influence of the domestic reservoirs on the temporal patter of water consumption of the residences. To mitigate these effects, modifications were proposed in the calibration and generation stages of the synthetic water demand generation series. In general, it was observed that these proposed modifications contributed to a more accurately reproduction of the observed series statistics by the OP and NSRP synthetic series, especially regarding the intensities and durations of the simulated demands. Although the modified versions of the NSRP and OP models presented similar performance in the reproduction of the means, variances and covariance of the observed series, the OP model reproduced in a more consistent way the observed daily consumed volumes.

Cette thèse se décompose en trois parties distinctes auxquelles s'ajoute une introduction. Dans un premier temps, nous nous intéressons à un test lisse d'ajustement à la famille de Pareto. Pour cela, nous proposons une statistique de test motivée par la théorie de LeCam sur la normalité asymptotique locale (LAN). Nous en établissons le comportement asymptotique sous l'hypothèse que l'échantillon provient d'une distribution de Pareto et sous des alternatives locales, nous plaçant ainsi dans le cadre LAN. Des simulations sont présentées afin d'étudier le comportement de la statistique de test à distance finie. Dans le chapitre suivant, nous nous plaçons dans le cadre de données censurées aléatoirement à droite. Nous proposons alors un estimateur des paramètres de la distribution de Pareto généralisée basé sur une première étape de l'algorithme de Newton-Raphson. Nous établissons la normalité asymptotique de cet estimateur. Par des simulations, nous illustrons son comportement à distance finie et le comparons à celui de l'estimateur du maximum de vraisemblance. Nous proposons enfin, dans un dernier chapitre, un modèle linéaire autorégressif adapté à la loi de Gumbel pour prendre en compte la dépendance dans les maxima. Nous établissons des propriétés théoriques de ce modèle et par simulations nous illustrons son comportement à distance finie. Enfin, comme des applications concrètes en sciences de l'atmosphère motivaient ce modèle, nous l'avons utilisé pour modéliser des maxima de dioxyde de carbone et de méthane.

Some financial problems as minimizing the shortfall risk when hedging in incomplete markets lead to problems belonging to test theory. This paper considers a generalization of the Neyman-Pearson lemma. With methods of convex duality we deduce the structure of an optimal randomized test when testing a compound hypothesis against a simple alternative. We give necessary and sufficient optimality conditions for the problem.

This thesis studies inference to the complier treatment effect denoted LATE. The standard approach is to base the inference on the two-stage least squares (2SLS) estimator and asymptotic Neyman inference, i.e., the t-test. The paper suggests a Fisher Randomization Test based on the t-test statistic as an alternative to the Neyman inference. Based on the setup with a randomized experiment with noncompliance, for which one can identify the LATE, I compare the two approaches in a Monte Carlo (MC) simulations. The results from the MC simulation is that the Fisher randomization test is not a valid alternative to the Neyman’s test as it has too low power.

No description available.

We provide a Frequentist-Bayesian hybrid test statistic in this dissertation for two testing problems. The first one is to design a test for the significant differences between non-parametric functions and the second one is to design a test allowing any departure of predictors of high dimensional X from constant. The implementation is also given in construction of the proposal test statistics for both problems. For the first testing problem, we consider the statistical difference among massive outcomes or signals to be of interest in many diverse fields including neurophysiology, imaging, engineering, and other related fields. However, such data often have nonlinear system, including to row/column patterns, having non-normal distribution, and other hard-to-identifying internal relationship, which lead to difficulties in testing the significance in difference between them for both unknown relationship and high-dimensionality. In this dissertation, we propose an Adaptive Bayes Sum Test capable of testing the significance between two nonlinear system basing on universal non-parametric mathematical decomposition/smoothing components. Our approach is developed from adapting the Bayes sum test statistic by Hart (2009). Any internal pattern is treated through Fourier transformation. Resampling techniques are applied to construct the empirical distribution of test statistic to reduce the effect of non-normal distribution. A simulation study suggests our approach performs better than the alternative method, the Adaptive Neyman Test by Fan and Lin (1998). The usefulness of our approach is demonstrated with an application in the identification of electronic chips as well as an application to test the change of pattern of precipitations. For the second testing problem, currently numerous statistical methods have been developed for analyzing high-dimensional data. These methods mainly focus on variable selection approach, but are limited for purpose of testing with high-dimensional data, and often are required to have explicit derivative likelihood functions. In this dissertation, we propose ``Hybrid Omnibus Test'' for high-dimensional data testing purpose with much less requirements. Our Hybrid Omnibus Test is developed under semi-parametric framework where likelihood function is no longer necessary. Our Hybrid Omnibus Test is a version of Freqentist-Bayesian hybrid score-type test for a functional generalized partial linear single index model, which has link being functional of predictors through a generalized partially linear single index. We propose an efficient score based on estimating equation to the mathematical difficulty in likelihood derivation and construct our Hybrid Omnibus Test. We compare our approach with a empirical likelihood ratio test and Bayesian inference based on Bayes factor using simulation study in terms of false positive rate and true positive rate. Our simulation results suggest that our approach outperforms in terms of false positive rate, true positive rate, and computation cost in high-dimensional case and low-dimensional case. The advantage of our approach is also demonstrated by published biological results with application to a genetic pathway data of type II diabetes. / Ph. D.

Les nouvelles réglementations concernant les usages de produits phytosanitaires et la prise en compte de l'environnement (pollution, biodiversité) en agriculture ont conduit à la mise au point de méthodes d'identification de plantes (culture et adventices) par une gestion spécifique des adventices par imagerie. Afin de disposer d'un outil performant permettant l'évaluation de ces méthodes d'identification reposant sur une analyse spatiale de la scène photographiée, un modèle de simulation de scènes agronomiques a été mis au point. Prenant en considération certaines caractéristiques agronomiques d'une parcelle cultivée, ce modèle permet de simuler une vérité terrain dont les paramètres - la spatialisation de la culture, le taux d'infestation, la distribution des adventices - sont contrôlés. La scène agronomique ainsi créée subit ensuite une transformation projective afin de simuler la prise de photographie et, ainsi, de prendre en compte tous les paramètres nécessaire à la création d'une image. Ce modèle a ensuite été validé à l'aide de comparaison statistique avec des données réelles. De nouveaux algorithmes spatiaux basés sur la Transformée de Hough et utilisant l'alignement en rang de la culture ont également été développés. Trois méthodes basées sur une analyse en composante connexe, une estimation de contours et une méthode probabiliste ont été mises en œuvre et exhaustivement évaluées à l'aide du modèle développé. Les résultats obtenus sont de très bonne qualité avec une classification correcte de la culture et des adventices supérieure à 90% et pouvant atteindre 98% dans certains cas. Enfin, pour ce modèle, une approche spectrale a également été explorée afin de dépasser les limitations imposées par les méthodes spatiales. Une extension 3D a été apportée à ce modèle afin de permettre la simulation de la réflectance bidirectionnelle (BRDF) des plantes et du sol à l'aide des modèles PROSPECT et SOILSPECT. La transformation d'une information spectrale en une information couleur RGB, la prise en compte de filtres optiques ou la création de données multispectrales sont également discutées.