B0290
Title: Functional data analysis in constrained spaces
Authors: John Aston - University of Cambridge (United Kingdom) [presenting]
Eardi Lila - University of Washington (United States)
Abstract: Traditional Functional Data has been analysed as 1-dimensional curves which are assumed to be i.i.d. However, in many settings, this is not the case. We will look at examples of functional data which is constrained to lie in certain spaces, which are often non-Euclidean. This limits usual estimation techniques and requires the development of new methods to take into account the geometry of the setting. We will illustrate the issues with data from neuroimaging which is both surface-based and is related to the covariance operator of a process.
