Discretization error for the maximum of a Gaussian field

Abstract

A Gaussian eld X dened on a square S of R 2 is considered. We assume that this eld is only observed at some points of a regular grid with spacing 1 n . We are interested in the normalized discretization error n 2 (M − M n ), with M the global maximum of X over S and M n the maximum of X over the observation grid. The density of the location of the maximum is given using Rice formulas and its regularity is studied. Joint densities with the value of the eld and the value of the second derivative are also given. Then, a kind of Slepian model is used to study the eld behavior around the unique point where the maximum is attained, called t * . We show that the normalized discretization error can be bounded by a quantity that converges in distribution to a uniform variable. The set where this uniform variable lies principally depends on the second derivative of the eld at t * . The bound is a function of this quantity which is approached by nite dierences in practice. The bound is applied both on simulated and real data. Real data are used in positioning by satellites systems quality assessment.

The harmonic mean formula for random processes

“…A simple application of Theorem 2.1 concerns the question of continuity of M(A , X), which has been investigated in various generalities in numerous contributions, see e.g., [8][9][10][11] and the excellent contribution [9], where the methodology is explained in great details.…”

The relation between wave asymmetry and particle orbits analysed by Slepian models

“…Examples of Slepian models in statistical theory are Gadrich & Adler (1993) on non-stationary processes, Baxevani & Wilson (2018) on prediction of extreme events in space over time, Baxevani, Podgórski & Rychlik (2003) on velocities of random surfaces, Azaïs & Chassan (2020) on statistical extreme value theory and Podgórski, Rychlik & Wallin (2015) on the Laplace moving average.…”

Statistics for Gaussian random fields with unknown location and scale using Lipschitz‐Killing curvatures

“…In this article, the images we consider are the excursion sets of the realization of a two‐dimensional Gaussian stationary and isotropic random field X above a given level u , that is, a black and white image indicating when the realization of X is above or below the level u . This work is part of a growing field of research, at the intersection of stochastic geometry and statistical analysis, which received an increasing attention in the recent years; see, for instance, Lachièze‐Rey (2019), Azaïs and Chassan (2020), Panigrahi et al (2019), Cheng (2016), Azaïs and Pham (2016), or Pham (2013). This literature investigates several geometrical stochastic objects and introduces consistent inference procedures to estimate them.…”