Title: Reconstruction of functional fragments
Authors: Marco Stefanucci - University of Rome - Sapienza (Italy) [presenting]
David Kraus - Masaryk University (Czech Republic)
Abstract: Often in the framework of functional data analysis the curves might be observed only on a subset of the whole domain, invalidating most of the existing statistical procedures. We propose a method able to handle this kind of data, in particular in the context of curve reconstruction, where the aim is to predict the unobserved part of each curve. More specifically, we study the role of regularization in the estimation of the reconstruction operator. The performances of different type of regularization are presented and the main differences with the standard estimator are discussed. The relevance of the proposed methodology is further pointed out through an application to a real dataset.