Théorème spectral | EPFL Graph Search

En mathématiques, et plus particulièrement en algèbre linéaire et en analyse fonctionnelle, on désigne par théorème spectral plusieurs énoncés affirmant, pour certains endomorphismes, l'existence de décompositions privilégiées, utilisant en particulier l'existence de sous-espaces propres. vignette|Une illustration du théorème spectral dans le cas fini : un ellipsoïde possède (en général) trois axes de symétrie orthogonaux (notés ici x, y et z). Le cas le plus élémentaire concerne les matrices symétriques représentant les formes quadratiques en dimension finie ; le théorème spectral correspondant, démontré par Karl Weierstrass en 1858, affirme que ces matrices sont toutes diagonalisables dans les réels, par l'intermédiaire d'un changement de base orthonormée ; un exemple de conséquence géométrique de ce résultat est l'existence, pour les quadriques non dégénérées, de trois axes de symétrie orthogonaux, les axes principaux, mais il a d'autres conséquences importantes dans des domaines m

A Euclidean Likelihood Estimator for Bivariate Tail Dependence

Boris Oumow, Michal Warchol

The spectral measure plays a key role in the statistical modeling of multivariate extremes. Estimation of the spectral measure is a complex issue, given the need to obey a certain moment condition. We propose a Euclidean likelihood-based estimator for the spectral measure which is simple and explicitly defined, with its expression being free of Lagrange multipliers. Our estimator is shown to have the same limit distribution as the maximum empirical likelihood estimator of J. H. J. Einmahl and J. Segers, Annals of Statistics 37(5B), 2953–2989 (2009). Numerical experiments suggest an overall good performance and identical behavior to the maximum empirical likelihood estimator. We illustrate the method in an extreme temperature data analysis.

Taylor & Francis Inc2013

The non-linear stochastic wave equation in high dimensions

Daniel Conus

The main topic of this thesis is the study of the non-linear stochastic wave equation in spatial dimension greater than 3 driven by spatially homogeneous Gaussian noise that is white in time. We are interested in questions of existence and uniqueness of solutions, as well as in properties of solutions, such as existence of high order moments and Hölder-continuity properties. The stochastic wave equation is formulated as an integral equation in which appear stochastic integrals with respect to martingale measures (in the sense of J.B. Walsh). Since, in dimensions greater than 3, the fundamental solution of the wave equation is neither a function nor a non-negative measure, but a general Schwartz distribution, we first develop an extension of the Dalang-Walsh stochastic integral that makes it possible to integrate a wide class of Schwartz distributions. This class contains the fundamental solution of the wave equation, under a hypothesis on the spectral measure of the noise that has already been used in the literature. With this extended stochastic integral, we establish existence of a square-integrable random-field solution to the non-linear stochastic wave equation in any dimension. Uniqueness of the solution is established within a specific class of processes. In the case of a fine multiplicative noise, we obtain a series representation of the solution and estimates on the p-th moments of the solution (p ≥ 1). From this, we deduce Hölder-continuity of the solution under standard assumptions. The Hölder exponent that we obtain is optimal. For the case of general multiplicative noise, we construct a framework for working with appropriate iterated stochastic integrals and then derive a truncated Itô-Taylor expansion for the solution of the stochastic wave equation. The convergence of this expansion remains an open problem, so we conclude with some remarks that suggest an Itô-Taylor series expansion for the solution.

EPFL2008

Exact simulation of max-stable processes

Clément Dombry, Sebastian Engelke

Max-stable processes play an important role as models for spatial extreme events. Their complex structure as the pointwise maximum over an infinite number of random functions makes their simulation difficult. Algorithms based on finite approximations are often inexact and computationally inefficient. We present a new algorithm for exact simulation of a max-stable process at a finite number of locations. It relies on the idea of simulating only the extremal functions, that is, those functions in the construction of a max-stable process that effectively contribute to the pointwise maximum. We further generalize the algorithm by Dieker & Mikosch (2015) for Brown-Resnick processes and use it for exact simulation via the spectral measure. We study the complexity of both algorithms, prove that our new approach via extremal functions is always more efficient, and provide closed-form expressions for their implementation that cover most popular models for max-stable processes and multivariate extreme value distributions. For simulation on dense grids, an adaptive design of the extremal function algorithm is proposed.

Oxford University Press2016

The non-linear stochastic wave equation in high dimensions

Daniel Conus

EPFL2008