Performance of the Gaussian model (GAUSS). The figure depicts the effect of the Gaussian scale parameter (θ in Eq. ) on the restricted log-likelihood (LL), the training population accuracy (rtrain), and the cross-validation accuracy (rpred) when predicting sets 5 or 6 in environment 1. For set 5 the restricted maximum likelihood solution for θ (θREML) = 0.5, and for set 6 θREML = 0.4. In both cases rtrain approached 1 as θ → 0, but the trends for rpred were different. For set 5 rpred exhibited an interior maximum near θREML, while for set 6 rpred increased monotonically with θ. Because GAUSS is approximately ridge regression (RR) when θ is large, the contrasting behavior in this figure illustrates why GAUSS had higher rpred than RR for set 5 but vice versa for set 6 (see Table 2).