Simultaneous inference for misaligned multivariate functional data

Multivariate functional data is a frequently observed data type, however there is only little literature exists on methods for misaligned multivariate functional data. In this seminar we will present a class of generally applicable models for handling multivariate functional data without the need of pre-processing. Inference of warp and amplitude effects is done simultaneously in an iterative procedure using Laplace approximation. We will apply it to two different datasets on human movement and finally compare it to other methods in a classification setup. contact: simone.vantini@polimi.it