Kolloquium: Norbert Henze (KIT): „Weighted L²-distances for Statistical Model Checks in Euclidean and Hilbert Spaces“

Weighted L²-distances for Statistical Model Checks in Euclidean and Hilbert Spaces – Vortragender: Norbert Henze, Karlsruher Institut für Technologie – Einladender: W. Stummer

Abstract: Weighted L²-distances can be used to build a flexible class of tools for testing hypotheses about the unknown distribution of the underlying random-data generating mechanism. Such hypotheses can concern both one-sample and multi-sample problems. The lecture presents typical test problems, with special emphasis on tests for multivariate normality and for Gaussianity in infinite-dimensional Hilbert spaces. It also highlights the mathematical challenges in connection with deriving asymptotic results as the sample size tends to infinity.