Spatial functional principal component analysis and its application in diagnostics

Abstract

In this thesis the nonparametric functional principal component analysis is extended to spatial data. The estimation is implemented in an R package and the implementation is evaluated comparing estimated eigenfunctions of realizations of a Wiener process to the theoretic ones. Furthermore, consistency results for the estimators are derived theoretically. Both, the one- and the two-dimensional method are used to evaluate several real data examples from the medical field, especially diagnostics.