A0434
Title: Change region detection on d-dimensional spheres
Authors: Di Su - London School of Economics (United Kingdom) [presenting]
Yining Chen - London School of Economics and Political Science (United Kingdom)
Tengyao Wang - London School of Economics (United Kingdom)
Abstract: While change point detection in time series data has been extensively studied, little attention has been given to its generalization to data observed on manifolds, where changes may occur within spatially complex regions with irregular boundaries, posing significant challenges. A new class of estimators is proposed to locate changes in the mean function of a signal-plus-noise model defined on d-dimensional spheres. This approach applies to scenarios with a single change region and multiple change regions. The convergence rate of the estimator is shown to depend on the VC dimension of the hypothesis class that characterizes the change regions. The results extend to data observed on d-dimensional manifolds under further assumptions. Simulations confirm the consistency of this approach, and the estimator's practical applicability is demonstrated through a global temperature dataset.
