Smooth fitting with a method for determining the regularization parameter under the genetic programming algorithm

This paper deals with the smooth fitting problem under the genetic programm ing (GP) algorithm. To reduce the computational cost required for evaluatin g the fitness value of GP trees, numerical weights of GP trees are estimate d by adopting both linear associative memories (LAM) and the Hook and Jeeve s (HJ) method. The quality of smooth fitting is critically dependent on the choice of the regularization parameter. So, we present a novel method for choosing the regularization parameter. Two numerical examples are given wit h the comparison of generalized cross-validation (GCV) B-splines. (C) 2001 Elsevier Science Inc. All rights reserved.