A0728
Title: Confidence regions for geometric features via extreme value distributions of Gaussian random fields on manifolds
Authors: Wanli Qiao - George Mason University (United States) [presenting]
Abstract: Geometric features, frequently expressed as level sets of functions in Euclidean spaces, often manifest as manifolds. Employing kernel estimators, establishing confidence regions for these features is commonly framed as discerning extreme value distributions of locally stationary Gaussian random fields along the level sets. The aim is to delve into the asymptotic formulations of these distributions and explore bootstrap techniques for their approximation. Furthermore, the application of these methodologies is demonstrated in constructing confidence regions for ridges.
