Using recursive algorithms for the efficient identification of smoothing spline ANOVA models

In this paper we present a unified discussion of different approaches to identification of smoothing spline ANOVA models: (i) the ¿classical¿ approach (in the line of Wahba, 1990; Gu, 2002; Storlie et al., 2009) and (ii) the so-called State-Dependent Regression (SDR) approach of Young (2001). The latter is a non-parametric approach which is very similar to smoothing splines and kernel regression approaches, but which is based on recursive filtering and smoothing estimation (the Kalman Filter combined with Fixed Interval Smoothing). We will show that the SDR approach can be optimally combined with the ¿classical¿ one, to obtain a more accurate and efficient estimation of smoothing splines ANOVA models to be applied for emulation purposes. We will also show that such an approach can compare favorably with kriging.

RATTO Marco;  PAGANO Andrea; 

2011-01-14

SPRINGER

JRC57205

1863-8171,   

http://www.springerlink.com/content/qu00g1803hv401h0/,    https://publications.jrc.ec.europa.eu/repository/handle/JRC57205,   

10.1007/s10182-010-0148-8,