We discuss the challenge in properly assessing the uncertainty of the estimates produced by the R package pffr, especially as it pertains to constructing confidence bands and computing p-values in functional linear models. We also present an approach that partially addresses some of these issues. Simulations are provided to help articulate these ideas.
ChoiHReimherrM (2016) A geometric approach to condence regions and bands for functional parameters. arXiv. Available from https://arxiv.org/abs/1607. 07771 (accessed on 23 November 2016).
2.
DegrasD (2011) Simultaneous confidence bands for nonparametric regression with functional data. Statistica Sinica, 21, 1735–65.
3.
SteinML (1999) Interpolation of Spatial Data: Some Theory for Krigging. New York: Springer.
