Abstract
It is shown in Huang and Chan (2013) that the local polynomial projection approach admits an equivalent mixed model formulation and they suggest a new smoothing approach using a combination of unpenalized polynomials and penalized trigonometric (Fourier) functions. We attempt to implement penalized Fourier regression by writing some R-functions in this thesis for easy and transparent usage of the methods. Our work is based on the book by Ruppert, Wand, and Carroll (2003) and improves some of their algorithms in the settings of univariate nonparametric regression, partial linear models, additive models, and nonparametric logistic and Poisson regression. We consider two forms of penalty for penalized Fourier regression, a quadratic penalty \alpha k^{2}, where k denotes the frequency of Fourier basis functions and \alpha\geqslant0, and the penalty estimated by REML mimicking that of penalized splines (Ruppert et al., 2003). Some data examples are used to illustrate our R-functions. A small simulation study is conducted to compare the penalized Fourier to spline approaches.