Halfspace depth for multivariate functional data
Speaker: Gerda Claeskens (KU Leuven)
Time: 3.00 PM
Date: Thursday Febuary 7th 2013
Location: Statistics Seminar Room L550B, 5th floor, Library Building
Abstract:
A multivariate depth (ordering from the center outward) for functional data is defined and studied. By the multivariate nature and by including a weight function, it acknowledges important characteristics of functional data, namely differences in the amount of local amplitude, shape and phase variation. Both population and finite sample versions are studied. The multivariate sample of curves may include warping functions, derivatives and integrals of the original curves for a better overall representation of the functional data via the depth.
Series: Statistics Seminar Series
Social Media Links