The problem of determining whether or not a theoretical model is an accurate representation of an empirically observed phenomenon is one of the most challenging in the empirical scientific investigation. The following study explores the problem of stochastic model validation. Special attention is devoted to the unusual two-peaked shape of the empirically observed distributions of the conditional on realised volatility financial returns. The application of statistical hypothesis testing and simulation techniques leads to the conclusion that the conditional on realised volatility returns are distributed with a specific previously undocumented distribution. The probability density that represents this distribution is derived, characterised and applied for validation of the financial model.

[pt] O ruído interferométrico é um fator limitante cada vez mais importante nos sistemas óticos, principalmente nas ligações de longa distância em redes óticas transparentes. O presente trabalho analisa modelos para este tipo de ruído, dando um tratamento matemático novo para o modelo não-gaussiano. A teoria matemática é desenvolvida em detalhes e comrigor. O modelo gaussiano foi usado a fim de fazer previsões quanto aos valores de chão da taxa de erro de bits. Os dois modelos foram simulados em computador e comparados com os testes realizados em laboratório e os resultados são apresentados. / [en] The interferometric noise is becoming a serious limiting factor in optical systems,notably on long distance connections in transparent optical networks. The present work analyzes models for this kind of noise, giving a new mathematical treatment to the non-gaussian model. The mathematical theory is developed in detail and rigorously. The gaussian model was used in order to make predictions relative to bit error rate floors. Both models were simulated in computer and compared with the tests made in laboratory and the results are presented. / [es] EL ruido interferométrico es un factor limitante cada vez más importante en los sistemas ópticos, principalmente en las llamadas a larga distancia en redes ópticas transparentes. EL presente trabajo analiza modelos para este tipo de ruido, dando un nuevo tratamiento matemático para el modelo no gausiano. La teoría matemática es desarrollada en detalles y con rigor. EL modelo gausiano fue usado para efectuar previsiones de las cotas inferiores de la tasa de error de bits. Los dos modelos fueron simulados en computador y comparados con las priuebas de laboratorio y se presentan los resultados.

[pt] O presente trabalho tem como objetivo estudar o comportamento de estruturas geotécnicas mediante o uso de Análise Limite Numérica. Para isto foi desenvolvido o programa GEOLIMA (GEOtechnical LIMit Analysis) com base na teoria de Análise Limite Numérica utilizando o Método de Elementos Finitos (MEF), considerando problemas bidimensionais e tridimensionais. Devido ao fato das propriedades do solo serem variáveis aleatórias, a Análise Não Determinística também foi considerada mediante o uso do Método Estatístico Linear e do Método de Monte Carlo. Inicialmente, são apresentados os fundamentos da teoria de Análise Limite Determinística e sua formulação mista pelo Método de Elementos Finitos. A seguir são apresentados os fundamentos de Análise Não Determinística, onde os métodos Estatístico Linear e Monte Carlo são descritos. As fases de desenvolvimento do GEOLIMA são descritas de forma resumida e a validação é feita mediante a comparação de resultados obtidos com soluções analíticas ou outras soluções. A seguir, uma aplicação em 2D é apresentada com a finalidade de ilustrar a Análise Limite Determinística e Não Determinística mediante o método Estatístico Linear e o método de Monte Carlo. Finalmente, duas aplicações em 3D são apresentadas: um problema relativo à frente de escavação de um túnel e um estudo de painéis de mineração. Os resultados deste trabalho indicam a viabilidade de usar Análise Limite Determinística e Não Determinística no estudo de problemas geotécnicos. / [en] The present work has the purpose of studying the behavior of geotechnical structures by means of numerical analysis. For this, program GEOLIMA (GEOtechnical LIMit Analysis) was developed based on the theory of Numerical Limit Analysis using the Finite Element Method (FEM), considering bidimensional and three-dimensional problems. Due to the fact that the properties of the ground are generally random variables, Non Deterministic Analysis was also considered by means of the Linear Statistical and the Monte Carlo Methods. Initially, the fundamentals of Deterministic Limit Analysis and its mixed formulation are presented. Then, the fundamentals of Non Deterministic Theory are presented, and the Linear Statistic and the Monte Carlo Methods are described. The development phases of GEOLIMA are briefly described. Its validation is made by comparing the results obtained with analytical solutions or other solutions. Following, a 2D application is made with the purpose of illustrating Deterministic and Non Deterministic Limit Analysis. Finally, two 3D applications are presented: a problem related to the excavation of a tunnel front and a problem related to mining panels. The results of this work indicate the viability of using Deterministic and Non Deterministic Limit Analysis in the study of geotechnical problems.

Statisticians are used to model interactions between variables on the basis of observed<p>data. In a lot of emerging fields, like bioinformatics, they are confronted with datasets<p>having thousands of variables, a lot of noise, non-linear dependencies and, only, tens of<p>samples. The detection of functional relationships, when such uncertainty is contained in<p>data, constitutes a major challenge.<p>Our work focuses on variable selection and network inference from datasets having<p>many variables and few samples (high variable-to-sample ratio), such as microarray data.<p>Variable selection is the topic of machine learning whose objective is to select, among a<p>set of input variables, those that lead to the best predictive model. The application of<p>variable selection methods to gene expression data allows, for example, to improve cancer<p>diagnosis and prognosis by identifying a new molecular signature of the disease. Network<p>inference consists in representing the dependencies between the variables of a dataset by<p>a graph. Hence, when applied to microarray data, network inference can reverse-engineer<p>the transcriptional regulatory network of cell in view of discovering new drug targets to<p>cure diseases.<p>In this work, two original tools are proposed MASSIVE (Matrix of Average Sub-Subset<p>Information for Variable Elimination) a new method of feature selection and MRNET (Minimum<p>Redundancy NETwork), a new algorithm of network inference. Both tools rely on<p>the computation of mutual information, an information-theoretic measure of dependency.<p>More precisely, MASSIVE and MRNET use approximations of the mutual information<p>between a subset of variables and a target variable based on combinations of mutual informations<p>between sub-subsets of variables and the target. The used approximations allow<p>to estimate a series of low variate densities instead of one large multivariate density. Low<p>variate densities are well-suited for dealing with high variable-to-sample ratio datasets,<p>since they are rather cheap in terms of computational cost and they do not require a large<p>amount of samples in order to be estimated accurately. Numerous experimental results<p>show the competitiveness of these new approaches. Finally, our thesis has led to a freely<p>available source code of MASSIVE and an open-source R and Bioconductor package of<p>network inference. / Doctorat en sciences, Spécialisation Informatique / info:eu-repo/semantics/nonPublished

Mit der vorgelegten Arbeit werden neue Beiträge zur Grundlagenforschung auf dem Gebiet der Grenzwertsätze der Wahrscheinlichkeitstheorie vorgelegt. Grenzwertsätze für Summen unabhängiger Zufallsgrößen nehmen unter den verschiedenartigsten Forschungsrichtungen der Wahrscheinlichkeitstheorie einen bedeutenden Platz ein und sind in der heutigen Zeit nicht mehr allein von theoretischem Interesse. In der Arbeit werden Ergebnisse zu neuere Problemstellungen aus der Summationstheorie unabhängiger Zufallsgrößen vorgestellt, die erstmalig in den fünfziger bzw. sechzger Jahren des 20. Jahrhunderts in der Literatur auftauchten und in den zurückliegenden Jahren mit großem Interesse untersucht wurden. International haben sich in der Theorie der Grenzwertsätze zwei Hauptrichtungen herauskristallisiert: Zum Einen die Fragen zur Konvergenzgeschwindigkeit, mit der eine Summenverteilungsfunktion gegen eine vorgegebene Grenzverteilungsfunktion konvergiert, und zum Anderen die Fragen nach einer Fehlerabschätzung zur Grenzverteilungsfunktion bei einem endlichen Summationsprozeß. Zuerst werden unbegrenz teilbare Grenzverteilungsfunktionen betrachtet und dann wird speziell die Normalverteilung als Grenzverteilung diskutiert. Als charakteristische Kenngrößen werden sowohl Momente oder einseitige Momente bzw. Pseudomomente benutzt. Die Fehlerabschätzungen werden sowohl als gleichmäßige wie auch ungleichmäßige Restgliedabschätzungen angegeben, einschließlich einer Beschreibung der dabei auftretenden absoluten Konstanten. Als Beweismethoden werden sowohl die Methode der charakteristischen Funktionen als auch direkte Methoden (Faltungsmethode) weiter ausgebaut. Für eine 1965 von Bikelis angegebene Fehlerabschätzung gelang es nun erstmalig, die auftretende absolute Konstante C mit C=114,667 numerisch abzuschätzen. Weiterhin werden in der Arbeit sogenannte Grenzwertsätze für mittlere Abweichungen studiert. Hier werden erstmalig auch Restgliedabschätzungen abgeleitet. Der in den letzten Jahren zum Beweis von Grenzwertsätzen eingeschlagene Weg über die Faltung von Verteilungsfunktionen erwies sich als bahnbrechend und bestimmte die Entwicklung sowohl der Theorie der Grenzwertsätze für mittlere und große Abweichungen als auch der Untersuchung zu den ungleichmäßigen Abschätzungen im zentralen Grenzwertsatz bedeutend. Die Faltungsmethode stellt in der vorliegenden Dissertationsschrift das hauptsächliche Beweisinstrument dar. Damit gelang es, eine Reihe neuer Ergebnisse zu erhalten und insbesondere mittels der elektronischen Datenverarbeitung neue numerische Resultate zu erhalten. / With the presented work new contributions to basic research in the field of limit theorems of probability theory are given. Limit theorems for sums of independent random variables taking on the most diverse lines of research in probability theory an important place in modern times and are no longer only of theoretical interest. In the work results are presented to newer problems on the summation theory of independent random variables, at first time in the fifties and sixties of the 20th Century appeared in the literature and have been studied in the past few years with great interest. International two main directions have emerged in the theory of limit theorems: Firstly, the questions on the convergence speed of a cumulative distribution function converges to a predetermined limit distribution function, and on the other hand the questions on an error estimate for the limit distribution function at a finite summation process. First indefinite divisible limit distribution functions are considered, then the normal distribution is specifically discussed as a limit distribution. As characteristic parameters both moments or one-sided moments or pseudo-moments are used. The error estimates are stated both in uniform as well as non-uniform residual bounds including a description of the occurring absolute constants. Both the method of characteristic functions as well as direct methods (convolution method) can be further expanded as proof methods. Now for the error estimate, 1965 given by Bikelis, was the first time to estimate the appearing absolute constant C with C = 114.667 numerically. Furthermore, in the work of so-called limit theorems for moderate deviations are studied. Here also remainder estimates are derived for the first time. In recent years to the proof of limit theorems the chosen way of the convolution of distribution functions proved to be groundbreaking and determined the development of both the theory of limit theorems for moderate and large deviations as well as the investigation into the nonuniform estimates in the central limit theorem significantly. The convolution method is in the present thesis, the main instrument of proof. Thus, it was possible to obtain a series of results and obtain new numerical results in particular by means of electronic data processing.

IntroStat was designed to meet the needs of students, primarily those in business, commerce and management, for a course in applied statistics. IntroSTAT is designed as a lecture-book. One of the aims is to maximize the time spent in explaining concepts and doing examples. The book is commonly used as part of first year courses into Statistics.

Title: Options under Stable Laws. Author: Andrea Karlová Department: Department of Probability and Mathematical Statistics Supervisor: Doc. Petr Volf, CSc. Abstract: Stable laws play a central role in the convergence problems of sums of independent random variables. In general, densities of stable laws are represented by special functions, and expressions via elementary functions are known only for a very few special cases. The convenient tool for investigating the properties of stable laws is provided by integral transformations. In particular, the Fourier transform and Mellin transform are greatly useful methods. We first discuss the Fourier transform and we give overview on the known results. Next we consider the Mellin transform and its applicability on the problem of the product of two independent random variables. We establish the density of the product of two independent stable random variables, discuss the properties of this product den- sity and give its representation in terms of power series and Fox's H-functions. The fourth chapter of this thesis is focused on the application of stable laws into option pricing. In particular, we generalize the model introduced by Louise Bachelier into stable laws. We establish the option pricing formulas under this model, which we refer to as the Lévy Flight...

No description available.

A literatura dispõe de métodos de diagnóstico para avaliar o ajuste de modelos lineares generalizados (MLGs) para medidas repetidas baseado em equações de estimação generalizada (EEG). No entanto, tais métodos não contemplam a distribuição binomial nem bancos de dados com observações faltantes. O presente trabalho generalizou os métodos já desenvolvidos para essas duas situações. Na construção de gráficos de probabilidade meio-normal com envelope simulado para a distribuição binomial, foi proposto um método para geração de variáveis aleatórias com distribuição marginal binomial correlacionadas, baseado na convolução de variáveis com distribuição de Poisson independentes. Os métodos de diagnóstico desenvolvidos foram aplicados em dados reais e simulados. / Literature provides diagnostic methods to assess the fit of generalized linear models (GLM) for repeated measures based on generalized estimating equations (GEE). Still, such methods do not include the binomial distribution or databases with missing observations. This work generalizes the methods already developed for these two situations. A method for generating random variables with correlated marginal binomial distributions based on convolution of independent Poisson random variables has been proposed for the construction of half-normal probability plots. The diagnostic methods developed were applied to real and simulated data.

Em uma detalhada revisão nós obtemos a lei do semi-círculo para a densidade de estados no ensemble gaussiano de Wigner. Também falamos sobre a analogia eletrostática de Dyson, enxergando os autovalores como cargas que se repelem no círculo unitário, mostrando que nesse caso a densidade de estados é uniforme. Em um contexto mais geral nós obtemos a lei do semicírculo, provando o teorema de Glivenko-Cantelli para variáveis fortemente correlacionadas usando um método combinatorial de contagem de trajetos, o que nos dá subsídios para falar em estabilidade da lei do semi-círculo. Também, nesta dissertação nós estudamos as funções de correlação nos ensembles gaussiano e circular, mostrando que sob um adequado reescalamento elas são idênticas. Outros ensembles nesta dissertação foram investigados usando o Método de Gram para o caso em que os autovalores são limitados em um intervalo. Computamos a densidade de estados para cada um desses ensembles. Mais precisamente no ensemble de Chebychev, os resultados foram obtidos analiticamente e nesse ensemble além da densidade de estados, também traçamos grá…cos da função de correlação truncada. / In a detailed review we obtain a semi-circle law for the density of states in theWigner’s Gaussian Ensemble. Also we talk about Dyson’s Analogy, seeing the eigenvalues like charges that repulse themselves in the unitary circle, showing that this case the density of states is uniform. In a more general context we obtain the semi-circle law, proving the Glivenko-Cantelli Theorem to strongly correlated variables, using a combinatorial method of Paths' Counting. Thus we are showing the stability of the semi-circle Law. Also, in this dissertation we study the correlation functions in the Gaussian and Circular ensembles showing that using the Gram's Method in the case that eigenvalues are limited in a interval. In these ensembles we computed the density of states. More precisely, in a Chebychev ensemble the results were obtained analytically. In this ensemble, we also obtain graphics of the truncated correlation function.