In the late 1990’s, the orthogonal series approach to nonparametric regression became increasingly popular; see Hart (1997), Ogden (1997), and Efromovich (1999). In particular, wavelets provide orthogonal series in which many of the series elements have small support. This has been found useful in some applications. We discuss the orthogonal series approach to nonparametric regression in the context of fitting linear models.
KeywordsRegression Tree Nonparametric Regression Ridge Regression Haar Wavelet Estimate Regression Coefficient
Unable to display preview. Download preview PDF.