No description available.

Les modèles de séries chronologiques avec variances conditionnellement hétéroscédastiques sont devenus quasi incontournables afin de modéliser les séries chronologiques dans le contexte des données financières. Dans beaucoup d'applications, vérifier l'existence d'une relation entre deux séries chronologiques représente un enjeu important. Dans ce mémoire, nous généralisons dans plusieurs directions et dans un cadre multivarié, la procédure dévéloppée par Cheung et Ng (1996) conçue pour examiner la causalité en variance dans le cas de deux séries univariées. Reposant sur le travail de El Himdi et Roy (1997) et Duchesne (2004), nous proposons un test basé sur les matrices de corrélation croisée des résidus standardisés carrés et des produits croisés de ces résidus. Sous l'hypothèse nulle de l'absence de causalité en variance, nous établissons que les statistiques de test convergent en distribution vers des variables aléatoires khi-carrées. Dans une deuxième approche, nous définissons comme dans Ling et Li (1997) une transformation des résidus pour chaque série résiduelle vectorielle. Les statistiques de test sont construites à partir des corrélations croisées de ces résidus transformés. Dans les deux approches, des statistiques de test pour les délais individuels sont proposées ainsi que des tests de type portemanteau. Cette méthodologie est également utilisée pour déterminer la direction de la causalité en variance. Les résultats de simulation montrent que les tests proposés offrent des propriétés empiriques satisfaisantes. Une application avec des données réelles est également présentée afin d'illustrer les méthodes / Time series models with conditionnaly heteroskedastic variances have become almost inevitable to model financial time series. In many applications, to confirm the existence of a relationship between two time series is very important. In this Master thesis, we generalize in several directions and in a multivariate framework, the method developed by Cheung and Ng (1996) designed to examine causality in variance in the case of two univariate series. Based on the work of El Himdi and Roy (1997) and Duchesne (2004), we propose a test based on residual cross-correlation matrices of squared residuals and cross-products of these residuals. Under the null hypothesis of no causality in variance, we establish that the test statistics converge in distribution to chi-square random variables. In a second approach, we define as in Ling and Li (1997) a transformation of the residuals for each residual time series. The test statistics are built from the cross-correlations of these transformed residuals. In both approaches, test statistics at individual lags are presented and also portmanteau-type test statistics. That methodology is also used to determine the direction of causality in variance. The simulation results show that the proposed tests provide satisfactory empirical properties. An application with real data is also presented to illustrate the methods

Les modèles de séries chronologiques avec variances conditionnellement hétéroscédastiques sont devenus quasi incontournables afin de modéliser les séries chronologiques dans le contexte des données financières. Dans beaucoup d'applications, vérifier l'existence d'une relation entre deux séries chronologiques représente un enjeu important. Dans ce mémoire, nous généralisons dans plusieurs directions et dans un cadre multivarié, la procédure dévéloppée par Cheung et Ng (1996) conçue pour examiner la causalité en variance dans le cas de deux séries univariées. Reposant sur le travail de El Himdi et Roy (1997) et Duchesne (2004), nous proposons un test basé sur les matrices de corrélation croisée des résidus standardisés carrés et des produits croisés de ces résidus. Sous l'hypothèse nulle de l'absence de causalité en variance, nous établissons que les statistiques de test convergent en distribution vers des variables aléatoires khi-carrées. Dans une deuxième approche, nous définissons comme dans Ling et Li (1997) une transformation des résidus pour chaque série résiduelle vectorielle. Les statistiques de test sont construites à partir des corrélations croisées de ces résidus transformés. Dans les deux approches, des statistiques de test pour les délais individuels sont proposées ainsi que des tests de type portemanteau. Cette méthodologie est également utilisée pour déterminer la direction de la causalité en variance. Les résultats de simulation montrent que les tests proposés offrent des propriétés empiriques satisfaisantes. Une application avec des données réelles est également présentée afin d'illustrer les méthodes / Time series models with conditionnaly heteroskedastic variances have become almost inevitable to model financial time series. In many applications, to confirm the existence of a relationship between two time series is very important. In this Master thesis, we generalize in several directions and in a multivariate framework, the method developed by Cheung and Ng (1996) designed to examine causality in variance in the case of two univariate series. Based on the work of El Himdi and Roy (1997) and Duchesne (2004), we propose a test based on residual cross-correlation matrices of squared residuals and cross-products of these residuals. Under the null hypothesis of no causality in variance, we establish that the test statistics converge in distribution to chi-square random variables. In a second approach, we define as in Ling and Li (1997) a transformation of the residuals for each residual time series. The test statistics are built from the cross-correlations of these transformed residuals. In both approaches, test statistics at individual lags are presented and also portmanteau-type test statistics. That methodology is also used to determine the direction of causality in variance. The simulation results show that the proposed tests provide satisfactory empirical properties. An application with real data is also presented to illustrate the methods

Dans cette thèse, nous présentons une étude rigoureuse des barres d'erreurs et des sensibilités de paramètres neutroniques (tels le k-eff) aux données nucléaires de base utilisées pour les calculer. Notre étude commence au niveau fondamental, i.e. les fichiers de données ENDF et leurs incertitudes, fournies sous la forme de matrices de variance/covariance, et leur traitement. Lorsqu'un calcul méthodique et consistant des sensibilités est consenti, nous montrons qu'une approche déterministe utilisant des formalismes bien connus est suffisante pour propager les incertitudes des bases de données avec un niveau de précision équivalent à celui des meilleurs outils disponibles sur le marché, comme les codes Monte-Carlo de référence. En appliquant notre méthodologie à trois exercices proposés par l'OCDE, dans le cadre des Benchmarks UACSA, nous donnons des informations, que nous espérons utiles, sur les processus physiques et les hypothèses sous-jacents aux formalismes déterministes utilisés dans cette étude.

Made available in DSpace on 2014-08-20T14:25:50Z (GMT). No. of bitstreams: 1 dissertacao_daniel_souza_cardoso.pdf: 4876863 bytes, checksum: f5480536e10753263d2ac710e5a3c660 (MD5) Previous issue date: 2010-03-24 / Whereas the State of Rio Grande do Sul (RS), have an economy directly dependent on agriculture and livestock sectors, which in different studies are reported as dependent on the variability of certain climatological elements, and the RS element water is regarded as fundamental. We conducted a study of the monthly total rainfall, to long 60 years (1948/2007), collected from 31 meteorological stations (EMs) and distributed geographically in the state. In the interest of contributing to the local society to predict possible shortages, and / or development of public policies for the use of water resources in urban and rural areas. In order to obtain a model that can provide an approximation of the behavior of the average rainfall for each of the six homogeneous regions, as defined in the literature (Marques, 2005), held has an harmonic analysis of the data previously adjusted to 30-day months. Before the analisys, the properties were checked for normality, homogeneity of variance and stationarity. The data tested for normality and homogeneity of variances, have not passed satisfactory in these tests and, hence, there was a transformation of data, generating new data sets that met the conditions of homogeneity of variance and normality. The relative increase in the trend, to long 60 years, ranged from 2,7 to 13,3% in the six homogeneous regions. Through harmonic analysis was obtained models that adequately represent the behavior of the average rainfall for the six homogeneous regions of RS, consisting of 3 or 4 sine waves, with one representation from 81 to 95% of the variability of the data. It was possible to found that some harmonics stood out, by have higher representation of the variability of the observed data, and the harmonic half stood out, in 50% of the models, and the harmonics quarterly and annual stood out, in 33,33% and 16,66% of, respectively. The models were tested to forecast, within the interval of 2003/2007, evolving in time according to the trend of time series of each region, being validated in residual analysis, by of residuals autocorrelation . Showing up as appropriate for forecast of future values. / Considerando que o Estado do Rio Grande do Sul (RS), possui uma economia diretamente dependente dos setores pecuário e agrícola, que em diferentes estudos são apontados como dependentes da variabilidade de alguns elementos climatológicos, e que no RS o elemento hídrico é considerado como fundamental. Realizou-se um estudo dos totais mensais de chuva ao longo de 60 anos (1948/2007), coletados de 31 estações meteorológicas (EMs) bem distribuídas, geograficamente, no Estado. Com o interesse de contribuir para a sociedade local, na previsão de possíveis racionamentos, e/ou na elaboração de políticas públicas para o uso dos recursos hídricos, nas áreas urbana e rural. Com o objetivo de obter um modelo, que possa apresentar uma aproximação do comportamento da precipitação pluvial média de cada uma das seis regiões homogêneas, já definidas na literatura (Marques, 2005), realizou-se uma análise harmônica dos dados previamente ajustados à meses de 30 dias. Antes da análise foram verificadas as propriedades de normalidade, homogeneidade de variâncias e estacionariedade. Os dados submetidos aos testes de normalidade, e de homogeneidade de variâncias, não obtiveram aprovação satisfatória nestes testes e, daí, realizou-se uma transformação de dados, gerando novos conjuntos de dados, que satisfizeram as condições de homogeneidade de variâncias e normalidade. O aumento relativo da tendência ao longo de 60 anos, variou de 2,7 a 13,3% nas seis regiões homogêneas. Através da análise harmônica obteve-se modelos que representam adequadamente o comportamento da precipitação pluvial média para as seis regiões homogêneas do RS, constituídos por 3 ou 4 ondas senoidais, apresentando uma representatividade de 81 a 95% da variabilidade dos dados. Foi possível constatar que alguns harmônicos destacaram-se por apresentar maior representatividade da variabilidade dos dados observados, sendo que o harmônico semestral destacou-se em 50% dos modelos, e que os harmônicos quadrimestral e anual destacaram-se em 33,33% e 16,66% destes, respectivamente. Os modelos foram testados para previsão, compreendida no intervalo de 2003/2007, evoluindo no tempo de acordo com a tendência das séries temporais de cada região, sendo validados na análise residual pela autocorrelação dos resíduos. Mostrando-se como adequados para previsão de valores futuros.

The dissertation thesis aims to establish a complex theoretical basis for a measurement of customer's profitability which can be easily used in the practice. The main reason for that is the absence of such publication in the current academic sphere. A combination of management accounting and knowledge of banking enables to achieve the aim. The thesis itself is closely linked to the banking practice. There are derived customer's profitability indicators as modifications of the popular RAROC in which a customer margin is used instead of net profit. There is designed a simplified way of operational cost allocation. Questionnaires and interviews with senior representatives of seven Czech banks helped to identify the most significant labor-intensity factors of corporate customers. The description of principal features of risk costs and derivation of formulas for profitability variances are also the part of the thesis. The theoretical part is followed by the practical one where a contemporary banking practice of corporate customer's profitability measurement on the Czech market is explored. It identified three weak points -- cost allocation, profitability variances and one-year horizon of a calculation. At the end, the theoretical basis is applied on an existing customer portfolio and the result is compared with a currently used customer's profitability measurement.

The unique problems encountered when analyzing weather data sets - that is, measurements taken while conducting a meteorological experiment- have forced statisticians to reconsider the conventional analysis methods and investigate permutation test procedures. The problems encountered when analyzing weather data sets are simulated for a Monte Carlo study, and the results of the parametric and permutation t-tests are compared with regard to significance level, power, and the average coilfidence interval length. Seven population distributions are considered - three are variations of the normal distribution, and the others the gamma, the lognormal, the rectangular and empirical distributions. The normal distribution contaminated with zero measurements is also simulated. In those simulated situations in which the variances are unequal, the permutation test procedure was performed using other test statistics, namely the Scheffe, Welch and Behrens-Fisher test statistics. / Mathematical Sciences / M. Sc. (Statistics)

State estimation plays an important role in modern power systems. The errors in the telemetered measurements and the connectivity information of the network will greatly contaminate the estimated system state. This dissertation provides solutions to suppress the influences of these errors. A two-stage state estimation algorithm has been utilized in topology error identification in the past decade. Chapter II discusses the implementation of this algorithm. A concise substation model is defined for this purpose. A friendly user interface that incorporates the two-stage algorithm into the conventional state estimator is developed. The performances of the two-stage state estimation algorithms rely on accurate determination of suspect substations. A comprehensive identification procedure is described in chapter III. In order to evaluate the proposed procedure, a topology error library is created. Several identification methods are comparatively tested using this library. A remote measurement calibration method is presented in chapter IV. The un-calibrated quantities can be related to the true values by the characteristic functions. The conventional state estimation algorithm is modified to include the parameters of these functions. Hence they can be estimated along with the system state variables and used to calibrate the measurements. The measurements taken at different time instants are utilized to minimize the influence of the random errors. A method for auto tuning of measurement weights in state estimation is described in chapter V. Two alternative ways to estimate the measurement random error variances are discussed. They are both tested on simulation data generated based on IEEE systems. Their performances are compared. A comprehensive solution, which contains an initialization process and a recursively updating process, is presented. Chapter VI investigates the errors introduced in the positive sequence state estimation due to the usual assumptions of having fully balanced bus loads/generations and continuously transposed transmission lines. Several tests are conducted using different assumptions regarding the availability of single and multi-phase measurements. It is demonstrated that incomplete metering of three-phase system quantities may lead to significant errors in the positive sequence state estimates for certain cases. A novel sequence domain three-phase state estimation algorithm is proposed to solve this problem.

Cette thèse s'intéresse à l'estimation statistique dans un modèle à variances isolées (modèle spike) de grande dimension. La théorie des matrices aléatoires permet de prendre en compte cette spécificité, puisque la plupart des résultats limites s'appliquent aux matrices dont la taille tend vers l'infini. Une part importante de ces résultats concerne la matrice de covariance empirique. Dans un premier temps, nous nous intéressons à l'estimation du nombre de facteurs/spikes. La différence de comportement des valeurs propres de la matrice de covariance empirique, selon que l'on considère celles correspondant aux spikes ou non, nous permet de construire un estimateur. Ce dernier correspond à la différence de deux valeurs propres consécutives ordonnées. Nous établissons la consistance de l'estimateur dans le cas où toutes les spikes sont distinctes, et le comparons à deux méthodes existantes à travers des simulations. L'estimateur dépend d'un seuil qui doit remplir certaines conditions. Dans la suite, nous étendons le résultat de consistance au cas d'égalité et améliorons l'estimateur en changeant de seuil. Dans un second temps, nous considérons les estimateurs du maximum de vraisemblance d'un modèle à facteurs strict à variance homoscédastique. En utilisant un théorème limite pour les statistiques spectrales linéaires, nous corrigeons l'estimateur de la variance commune en grande dimension en donnant l'expression de son biais et en établissant sa loi limite. Nous présentons une version corrigée du test du rapport de vraisemblance d'adéquation à un modèle à facteurs. Finalement, nous construisons un test d'égalité de deux spikes.

The management and administration of legally defined marine boundaries in Australia is subject to a variety of political, legal and technical challenges. The purpose of this thesis is to address three of the technical challenges faced in the implementation of marine boundaries which cannot be dealt with by applying conventional land cadastre and land administration principles. The three challenges that are identified and addressed are (i) marine boundary delimitation and positioning uncertainty, (ii) the construction and maintenance of four dimensional marine parcels, and (iii) the modelling and management of marine boundary uncertainty metadata.