The subject of the thesis is the derivation of non-local diffusion equations from kinetic models with heavy-tailed equilibrium in velocity. We are particularly interested in confining the kinetic equations and developing methods that allow us, from the confined kinetic models, to derive confined versions of non-local diffusion equations.

Made available in DSpace on 2016-03-15T19:30:52Z (GMT). No. of bitstreams: 1 Jose Antonio Chavez Orlotegui.pdf: 1008123 bytes, checksum: 851740c2da263d053e73b2b6c4ab1a2a (MD5) Previous issue date: 2011-11-09 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The objective is to price agricultural areas and quantify the influence of prices of agricultural products in the value of rural enterprises. The assessment takes into account that the value of the area is linked to the Present Value of future cash flow from exploration and the existence of flexibility associated with land tenure that can be evaluated by the method of real options. Crops evaluated in the study will be the corn and soybean. The evolution of crop prices will be modeled on the CIR process (Cox, Ingersoll, Ross). The price of land will be calculated from a partial differential equation (PDE). This is addressed from Monte Carlo simulations. The results show that the price of choice, for both cultures, it is more than 40% of the land. The value of land for corn, the 30-year horizon, is greater than of soybeans. / O objetivo do estudo consiste em precificar áreas agrícolas e quantificar a influência dos preços de produtos agrícolas no valor da empresa rural. A avaliação considera que o valor da área está vinculado ao Valor Presente dos Fluxos de Caixa Futuros, da exploração e da existência de certa flexibilidade associada à posse da terra que pode ser avaliada por meio do método de opções reais. As culturas avaliadas no estudo serão a Soja e o Milho. A evolução dos preços das culturas será modelada a partir do processo CIR (Cox, Ingersoll, Ross). O preço da terra será calculado a partir de uma equação diferencial parcial (EDP). Esta é resolvida a partir de simulações de Monte Carlo. Os resultados mostram que o preço da opção, para ambas as culturas, vale mais de quarenta por cento do valor da terra. O valor da terra para o milho, no horizonte de trinta anos, é maior que o da soja.

Masters of Commerce / If share price fluctuations follow a simple random walk then it implies that forecasting models based on historical growth rates have little ability to forecast acceptable share price movements over a certain period. The simple random walk description of share price dynamics is obtained when a large number of investors have equal probability to buy or sell based on their own opinion. This simple random walk description of the stock market is in essence the Efficient Market Hypothesis, EMT. EMT is the central concept around which financial modelling is based which includes the Black-Scholes model and other important theoretical underpinnings of capital market theory like mean-variance portfolio selection, arbitrage pricing theory (APT), security market line and capital asset pricing model (CAPM). These theories, which postulates that risk can be reduced to zero sets the foundation for option pricing and is a key component in financial software packages used for pricing and forecasting in the financial industry. The model used by Black and Scholes and other models mentioned above are Gaussian, i.e. they exhibit a random nature. This Gaussian property and the existence of expected returns and continuous time paths (also Gaussian properties) allow the use of stochastic calculus to solve complex Black- Scholes models. However, if the markets are not Gaussian then the idea that risk can be. (educed to zero can lead to a misleading and potentially disastrous sense of security on the financial markets. This study project test the null hypothesis - share prices on the JSE follow a random walk - by means of graphical techniques such as symmetry plots and Quantile-Quantile plots to analyse the test distributions. In both graphical techniques evidence for the rejection of normality was found. Evidenceleading to the rejection of the hypothesis was also found through nonparametric or distribution free methods at a 1% level of significance for Anderson-Darling and Runs test.

Diese Dissertation betrachtet das Problem der nichtparametrischen Schätzung der Diffusionskoeffizienten eines ein-dimensionalen und zeitlich homogenen Itô-Diffusionsprozesses. Dabei werden verschiedene diskrete Sampling Regimes untersucht. Im ersten Teil zeigen wir, dass eine Variante des von Gobet, Hoffmann und Reiß konstruierten Niedrigfrequenz-Schätzers auch im Fall von zufälligen Beobachtungszeiten verwendet werden kann. Wir beweisen, dass der Schätzer optimal im Minimaxsinn und adaptiv bezüglich der Verteilung der Beobachtungszeiten ist. Außerdam wenden wir die Lepski Methode an um einen Schätzer zu erhalten, der zusätzlich adaptiv bezüglich der Sobolev-Glattheit des Drift- und Volatilitätskoeffizienten ist. Im zweiten Teil betrachten wir das Problem der Volatilitätsschätzung für äquidistante Beobachtungen. Im Fall eines stationären Prozesses, mit kompaktem Zustandsraum, erhalten wir einen Schätzer, der sowohl bei hochfrequenten als auch bei niedrigfrequenten Beobachtungen die optimale Minimaxrate erreicht. Die Konstruktion des Schätzers beruht auf spektralen Methoden. Im Fall von niedrigfrequenten Beobachtungen ist die Analyse des Schätzers ähnlich wie diejenige in der Arbeit von Gobet, Hoffmann und Reiß. Im hochfrequenten Fall hingegen finden wir die Konvergenzraten durch lokale Mittelwertbildung und stellen daubt eine Verbindung zum Hochfrequenzschätzer von Florens-Zmirou her. In der Analyse unseres universalen Schätzers benötigen wir scharfe obere Schranken für den Schätzfehler von Funktionalen der Occupation time für unstetige Funktionen. Wir untersuchen eine auf Riemannsummen basierende Approximation der Occupation time eines stationären, reversiblen Markov-Prozesses und leiten obere Schranken für den quadratischen Fehler her. Im Fall von Diffusionsprozessen erhalten wir Konvergenzraten für Sobolev Funktionen. / In this thesis, we consider the problem of nonparametric estimation of the diffusion coefficients of a scalar time-homogeneous Itô diffusion process from discrete observations under various sampling assumptions. In the first part, the low-frequency estimation method proposed by Gobet, Hoffmann and Reiß is modified to cover the case of random sampling times. The estimator is shown to be optimal in the minimax sense and adaptive to the sampling distribution. Moreover, Lepski''s method is applied to adapt to the unknown Sobolev smoothness of the drift and volatility coefficients. In the second part, we address the problem of volatility estimation from equidistant observations without a predefined frequency regime. In the case of a stationary diffusion with compact state space and boundary reflection, we introduce a universal estimator that attains the minimax optimal convergence rates for both low and high-frequency observations. Being based on the spectral method, the low-frequency analysis is similar to the study conducted by Gobet, Hoffmann and Reiß. On the other hand, the derivation of the convergence rates in the high-frequency regime requires local averaging of the low-frequency estimator, which makes it mimic the behaviour of the classical high-frequency estimator introduced by Florens-Zmirou. The analysis of the universal estimator requires tight upper bounds on the estimation error of the occupation time functional for non-continuous functions. In the third part of the thesis, we thus consider the Riemann sum approximation of the occupation time functional of a stationary, time-reversible Markov process. Upper bounds on the squared mean estimation error are provided. In the case of diffusion processes, convergence rates for Sobolev regular functions are obtained.

This thesis is concerned with the issue of extinction of populations composed of different types of individuals, and their behavior before extinction and in case of a very late extinction. We approach this question firstly from a strictly probabilistic viewpoint, and secondly from the standpoint of risk analysis related to the extinction of a particular model of population dynamics. In this context we propose several statistical tools. The population size is modeled by a branching process, which is either a continuous-time multitype Bienaymé-Galton-Watson process (BGWc), or its continuous-state counterpart, the multitype Feller diffusion process. We are interested in different kinds of conditioning on non-extinction, and in the associated equilibrium states. These ways of conditioning have been widely studied in the monotype case. However the literature on multitype processes is much less extensive, and there is no systematic work establishing connections between the results for BGWc processes and those for Feller diffusion processes. In the first part of this thesis, we investigate the behavior of the population before its extinction by conditioning the associated branching process X_t on non-extinction (X_t≠0), or more generally on non-extinction in a near future 0≤θ<∞ (X_{t+θ}≠0), and by letting t tend to infinity. We prove the result, new in the multitype framework and for θ>0, that this limit exists and is non-degenerate. This reflects a stationary behavior for the dynamics of the population conditioned on non-extinction, and provides a generalization of the so-called Yaglom limit, corresponding to the case θ=0. In a second step we study the behavior of the population in case of a very late extinction, obtained as the limit when θ tends to infinity of the process conditioned by X_{t+θ}≠0. The resulting conditioned process is a known object in the monotype case (sometimes referred to as Q-process), and has also been studied when X_t is a multitype Feller diffusion process. We investigate the not yet considered case where X_t is a multitype BGWc process and prove the existence of the associated Q-process. In addition, we examine its properties, including the asymptotic ones, and propose several interpretations of the process. Finally, we are interested in interchanging the limits in t and θ, as well as in the not yet studied commutativity of these limits with respect to the high-density-type relationship between BGWc processes and Feller processes. We prove an original and exhaustive list of all possible exchanges of limit (long-time limit in t, increasing delay of extinction θ, diffusion limit). The second part of this work is devoted to the risk analysis related both to the extinction of a population and to its very late extinction. We consider a branching population model (arising notably in the epidemiological context) for which a parameter related to the first moments of the offspring distribution is unknown. We build several estimators adapted to different stages of evolution of the population (phase growth, decay phase, and decay phase when extinction is expected very late), and prove moreover their asymptotic properties (consistency, normality). In particular, we build a least squares estimator adapted to the Q-process, allowing a prediction of the population development in the case of a very late extinction. This would correspond to the best or to the worst-case scenario, depending on whether the population is threatened or invasive. These tools enable us to study the extinction phase of the Bovine Spongiform Encephalopathy epidemic in Great Britain, for which we estimate the infection parameter corresponding to a possible source of horizontal infection persisting after the removal in 1988 of the major route of infection (meat and bone meal). This allows us to predict the evolution of the spread of the disease, including the year of extinction, the number of future cases and the number of infected animals. In particular, we produce a very fine analysis of the evolution of the epidemic in the unlikely event of a very late extinction. / Diese Arbeit befasst sich mit der Frage des Aussterbens von Populationen verschiedener Typen von Individuen. Uns interessiert das Verhalten vor dem Aussterben sowie insbesondere im Falle eines sehr späten Aussterbens. Wir untersuchen diese Fragestellung zum einen von einer rein wahrscheinlichkeitstheoretischen Sicht und zum anderen vom Standpunkt der Risikoanalyse aus, welche im Zusammenhang mit dem Aussterben eines bestimmten Modells der Populationsdynamik steht. In diesem Kontext schlagen wir mehrere statistische Werkzeuge vor. Die Populationsgröße wird entweder durch einen zeitkontinuierlichen mehrtyp-Bienaymé-Galton-Watson Verzweigungsprozess (BGWc) oder durch sein Analogon mit kontinuierlichem Zustandsraum, den Feller Diffusionsprozess, modelliert. Wir interessieren uns für die unterschiedlichen Arten auf Überleben zu bedingen sowie für die hierbei auftretenden Gleichgewichtszustände. Diese Bedingungen wurden bereits weitreichend im Falle eines einzelnen Typen studiert. Im Kontext von mehrtyp-Verzweigungsprozessen hingegen ist die Literatur weniger umfangreich und es gibt keine systematischen Arbeiten, welche die Ergebnisse von BGWc Prozessen mit denen der Feller Diffusionsprozesse verbinden. Wir versuchen hiermit diese Lücke zu schliessen. Im ersten Teil dieser Arbeit untersuchen wir das Verhalten von Populationen vor ihrem Aussterben, indem wir das zeitasymptotysche Verhalten des auf Überleben bedingten zugehörigen Verzweigungsprozesses (X_t|X_t≠0)_t betrachten (oder allgemeiner auf Überleben in naher Zukunft 0≤θ<∞, (X_t|X_{t+θ}≠0)_t). Wir beweisen das Ergebnis, neuartig im mehrtypen Rahmen und für θ>0, dass dieser Grenzwert existiert und nicht-degeneriert ist. Dies spiegelt ein stationäres Verhalten für auf Überleben bedingte Bevölkerungsdynamiken wider und liefert eine Verallgemeinerung des sogenannten Yaglom Grenzwertes (welcher dem Fall θ=0 entspricht). In einem zweiten Schritt studieren wir das Verhalten der Populationen im Falle eines sehr späten Aussterbens, welches wir durch den Grenzübergang auf θ→∞ erhalten. Der resultierende Grenzwertprozess ist ein bekanntes Objekt im eintypen Fall (oftmals als Q-Prozess bezeichnet) und wurde ebenfalls im Fall von mehrtyp-Feller-Diffusionsprozessen studiert. Wir untersuchen den bisher nicht betrachteten Fall, in dem X_t ein mehrtyp-BGWc Prozess ist und beweisen die Existenz des zugehörigen Q-Prozesses. Darüber hinaus untersuchen wir seine Eigenschaften einschließlich der asymptotischen und weisen auf mehrere Auslegungen hin. Schließlich interessieren wir uns für die Austauschbarkeit der Grenzwerte in t und θ, und die Vertauschbarkeit dieser Grenzwerte in Bezug auf die Beziehung zwischen BGWc und Feller Prozessen. Wir beweisen die Durchführbarkeit aller möglichen Grenzwertvertauschungen (Langzeitverhalten, wachsende Aussterbeverzögerung, Diffusionslimit). Der zweite Teil dieser Arbeit ist der Risikoanalyse in Bezug auf das Aussterben und das sehr späte Aussterben von Populationen gewidmet. Wir untersuchen ein Modell einer verzweigten Bevölkerung (welches vor allem im epidemiologischen Rahmen erscheint), für welche ein Parameter der Reproduktionsverteilung unbekannt ist. Wir konstruieren Schätzer, die an die jeweiligen Stufen der Evolution adaptiert sind (Wachstumsphase, Verfallphase sowie die Verfallphase, wenn das Aussterben sehr spät erwartet wird), und beweisen zudem deren asymptotische Eigenschaften (Konsistenz, Normalverteiltheit). Im Besonderen bauen wir einen für Q-Prozesse adaptierten kleinste-Quadrate-Schätzer, der eine Vorhersage der Bevölkerungsentwicklung im Fall eines sehr späten Aussterbens erlaubt. Dies entspricht dem Best- oder Worst-Case-Szenario, abhängig davon, ob die Bevölkerung bedroht oder invasiv ist. Diese Instrumente ermöglichen uns die Betrachtung der Aussterbensphase der Bovinen spongiformen Enzephalopathie Epidemie in Großbritannien. Wir schätzen den Infektionsparameter in Bezug auf mögliche bestehende Quellen der horizontalen Infektion nach der Beseitigung des primären Infektionsweges (Tiermehl) im Jahr 1988. Dies ermöglicht uns eine Vorhersage des Verlaufes der Krankheit inklusive des Jahres des Aussterbens, der Anzahl von zukünftigen Fällen sowie der Anzahl infizierter Tiere. Insbesondere ermöglicht es uns die Erstellung einer sehr detaillierten Analyse des Epidemieverlaufs im unwahrscheinlichen Fall eines sehr späten Aussterbens.

In 1988, Owen introduced empirical likelihood as a nonparametric method for constructing confidence intervals and regions. Since then, empirical likelihood has been studied extensively in the literature due to its generality and effectiveness. It is well known that empirical likelihood has several attractive advantages comparing to its competitors such as bootstrap: determining the shape of confidence regions automatically using only the data; straightforwardly incorporating side information expressed through constraints; being Bartlett correctable. The main part of this thesis extends the empirical likelihood method to several interesting and important statistical inference situations. This thesis has four components. The first component (Chapter II) proposes a smoothed jackknife empirical likelihood method to construct confidence intervals for the receiver operating characteristic (ROC) curve in order to overcome the computational difficulty when we have nonlinear constrains in the maximization problem. The second component (Chapter III and IV) proposes smoothed empirical likelihood methods to obtain interval estimation for the conditional Value-at-Risk with the volatility model being an ARCH/GARCH model and a nonparametric regression respectively, which have applications in financial risk management. The third component(Chapter V) derives the empirical likelihood for the intermediate quantiles, which plays an important role in the statistics of extremes. Finally, the fourth component (Chapter VI and VII) presents two additional results: in Chapter VI, we present an interesting result by showing that, when the third moment is infinity, we may prefer the Student's t-statistic to the sample mean standardized by the true standard deviation; in Chapter VII, we present a method for testing a subset of parameters for a given parametric model of stationary processes.

We propose a simple and general method to obtain the boundary crossing probability for Brownian motion. This method can be easily extended to higher dimensional of Brownian motion. It also covers certain classes of stochastic processes associated with Brownian motion. The basic idea of the method is based on being able to construct a nite Markov chain such that the boundary crossing probability of Brownian motion is obtained as the limiting probability of the nite Markov chain entering a set of absorbing states induced by the boundary. Numerical results are given to illustrate our method.

We propose a simple and general method to obtain the boundary crossing probability for Brownian motion. This method can be easily extended to higher dimensional of Brownian motion. It also covers certain classes of stochastic processes associated with Brownian motion. The basic idea of the method is based on being able to construct a nite Markov chain such that the boundary crossing probability of Brownian motion is obtained as the limiting probability of the nite Markov chain entering a set of absorbing states induced by the boundary. Numerical results are given to illustrate our method.

Nous nous intéressons à l'équivalence asymptotique, au sens de Le Cam, entre différents modèles statistiques. Plus précisément, nous avons exploré le cas de modèles statistiques associés à l'observation discrète de processus à sauts ou de diffusions unidimensionnelles, ainsi que des modèles à densité plus classiques.Ci-dessous, nous présentons brièvement les différents chapitres de la thèse.Nous commençons par présenter tous nos résultats dans un premier chapitre introductif. Ensuite, dans le Chapitre 2 nous rappelons les points clés de la théorie de Le Cam sur les expériences statistiques en se plaçant dans un contexte non paramétrique.Les Chapitres 3 et 4 traitent de l'équivalence asymptotique pour des modèles statistiques associés à l'observation discrète (haute fréquence) de processus à sauts. Dans un premier temps nous nous focalisons sur un problème d'équivalence en ce qui concerne l'estimation de la dérive, supposée appartenir à une certaine classe fonctionnelle. Il s'avère (Chapitre 3) qu'il y a une équivalence asymptotique, en ce qui concerne l'estimation de la dérive, entre le modèle statistique associé à l'observation discrète d'un processus additif $X$ et le modèle statistique gaussien associé à l'observation discrète de la partie continue de $X$.Dans un deuxième temps, nous nous sommes intéressés au problème de l'estimation non paramétrique de la densité de Lévy $f$ relative à un processus de Lévy à sauts purs, $Y$. Le Chapitre 4 illustre l'équivalence asymptotique, en ce qui concerne l'estimation de $f$, entre le modèle statistique associé à l'observation discrète de $Y$ et un certain modèle de bruit blanc gaussien ayant $sqrt f$ comme dérive.Le Chapitre 5 présente une extension d'un résultat bien connu sur l'équivalence asymptotique entre un modèle à densité et un modèle de bruit blanc gaussien.Le Chapitre 6 étudie l'équivalence asymptotique entre un modèle de diffusion scalaire avec une dérive inconnue et un coefficient de diffusion qui tend vers zéro et le schéma d'Euler correspondant.Dans le Chapitre 7 nous présentons une majoration en distance $L_1$ entre les lois de processus additifs.Le Chapitre 8 est consacré aux conclusions et discute des extensions possibles des travaux de thèse. / The subject of this Ph.D. thesis is the asymptotic equivalence, in the Le Cam sense, between different statistical models. Specifically, we explore the case of statistical models associated with the discrete observation of jump processes or diffusion processes as well as more classical density models.Below, we briefly introduce the different chapters of this dissertation.We begin by presenting our results in a first introductory chapter. Then, in Chapter 2, we recall the key points of the Le Cam theory on statistical experiences focusing on a nonparametric context.Chapters 3 and 4 deal with asymptotic equivalences for statistical models associated with discrete observation (high frequency) of jump processes. First, we focus on an equivalence problem regarding the estimation of the drift, assumed to belong to a certain functional class. It turns out (Chapter 3) that there is an asymptotic equivalence, for what concerns the estimation of the drift, between the statistical model associated with the discrete observation of an additive process $X$ and the Gaussian statistical model associated with the discrete observation of the continuous part of $X$. Then we study the problem of nonparametric density estimation for the Lévy density $f$ of a pure jump Lévy process $Y$. Chapter 4 illustrates the asymptotic equivalence, for what concerns the estimation of $f$, between the statistical model associated with the discrete observation of $Y$ and a certain Gaussian white noise model having $sqrt f$ as drift.In Chapter 5 we present an extension of the well-known asymptotic equivalence between density estimation experiments and a Gaussian white noise model.Chapter 6 describes the asymptotic equivalence between a scalar diffusion model with unknown drift and with diffusion coefficient tending to zero and the corresponding Euler scheme. In Chapter 7 we present a bound for the $L_1$ distance between the laws of additive processes.Chapter 8 is devoted to conclusions and discusses possible extensions of the results of this thesis.

Controlled linear stochastic evolution equations driven by Lévy processes are studied in the Hilbert space setting. The control operator may be unbounded which makes the results obtained in the abstract setting applicable to parabolic SPDEs with boundary or point control. The first part contains some preliminary technical results, notably a version of Itô formula which is applicable to weak/mild solutions of controlled equations. In the second part, the ergodic control problem is solved: The feedback form of the optimal control and the formula for the optimal cost are found. The control problem is solved in the mean-value sense and, under selective conditions, in the pathwise sense. As examples, various parabolic type controlled SPDEs are studied. 1