Menu Content/Inhalt
Seminars Print
previous year previous month next month next year
See by year See by month See by week See Today Search Jump to month
Geerda Claeskens, KUL Print
Friday, 24 September 2010, 14:30 - 15:30

Multiresolution warping of functional data

Gerda Claeskens, OR and Katholieke Universiteit Leuven

Abstract: Warping is an approach that reduces phase variability in functional observations. The alignment of the curves takes place by applying a smooth bijection to the functional argument. We propose a natural representation of warping functions in terms of a new type of elementary functions named `warplets', which are combined into the warping function by composition. The inverse warping function is trivial and explicit to obtain. A sequential Bayesian estimation strategy is introduced and implemented in a R library, which fits a series of models and transfers the posterior of the previous fit to the prior of the next fit. The information contained in the warping action is further actively used for clustering a set of functional data.
This is joint work with Leen Slaets and Bernard W. Silverman.

Location: Jacqueline Bottemanne,
Contact: Plaine NO9 salle des profs