ABSTRACT
As discussed in Sec. 5.1.1, the minimization of the fitness function in a statistical regression problem should be nested in a way that, if applicable, the inner minimization is carried out analytically. Here, we derive the form of the fitness function in the outer minimization for the parametric and non-parametric regression examples in Sec. 1.3.1 and Sec. 1.3.2 respectively. In each case, we start with the least squares function, L S, and minimize it analytically over the parameters that appear linearly. The resulting function, l S, is the one that is then minimized numerically.
