|
Abstract:
|
Search for hotspots have a wide range of applications, including in medical imaging, ecological monitoring over very large landscapes, analysis of ozone fields on a global scale, and others. We will consider spatial observations that are Gaussian subordinated, with marginal distributions that are location dependent and have spatial autocorrelations. The aim is to identify locations where a derivative of the mean surface locally exceeds a given threshold. We will discuss asymptotic results under various correlation types, as well as applications to some large-scale spatial environmental data. This is joint work with Gabrielle Moser, Statistics Lab, Swiss Federal Research Institute WSL.
|
