Title: Some theoretic results on functional data analysis with penalized splines
Authors: Luo Xiao - North Carolina State University (United States) [presenting]
Abstract: Penalized spline methods are popular for analyzing functional data. However, the theoretical foundation of penalized splines in functional data analysis is largely unknown. We attempt to fill the theoretical gap and introduce some new theoretical results in the estimation of the mean and covariance functions via penalized splines. In particular, we show that penalized splines are rate optimal and have two-type asymptotics similar to previous results for univariate penalized splines in the context of nonparametric regression.