|
Activity Number:
|
8
- Geometric Statistics for Complex Data
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Royal Statistical Society
|
|
Abstract #316760
|
|
|
Title:
|
The Shape Space of Functional Data
|
|
Author(s):
|
Karthik Bharath* and Ian H Jermyn
|
|
Companies:
|
University of Nottingham and Durham University
|
|
Keywords:
|
Time warping;
local extrema;
maximal invariant;
principle bundle
|
|
Abstract:
|
Statistical modelling of functional data in the presence of phase variation is influenced by alignment or registration procedures. The use of shape of functional data through group invariant methods presents a viable approach to carry out registration agnostic modelling. In this talk I will discuss the geometric and topological structure of the shape space of functional data and identify shape features which enable a suitable phase-invariant dimension reduction mechanism. Implications of the structure on identifiability of models and behaviour of kernel-based estimators will be discussed with examples.
|
Authors who are presenting talks have a * after their name.
