ABSTRACT
We present in this chapter the class of kernel-based local smoothing methods for nonparametric curve estimation with unstructured nonparametric regression models. These smoothing methods form a series of building blocks for constructing smoothing estimators when the nonparametric models are subject to structural constraints. In addition, as discussed in Section 2.3, these smoothing estimators can also be used in conjunction with parametric estimation procedures in semiparametric partially linear regression models.
