Table 2.

The mean of $\overline{R}$ and $\sqrt{Var(R)}$ computed from the predictions of networks $ne{t}_A$ and $ne{t}_B$.

	$\phi$	k	T
${\mathit{net}}_{\mathit{A}}$
Training set
$\overline{R}$	− 5.27e−04	− 8.63e−04	− 1.65e−04
$\sqrt{Var(R)}$	9.40e−03	7.54e−02	2.47e−03
Validation set
$\overline{R}$	− 5.83e−04	− 8.88e−04	− 1.37e−04
$\sqrt{Var(R)}$	9.33e−03	5.55e−02	2.45e−03
${\mathit{net}}_{\mathit{B}}$
Training set
$\overline{R}$	− 2.64e−04	− 1.91e−04	2.81e−06
$\sqrt{Var(R)}$	7.44e−03	3.35e−02	2.01e−03
Validation set
Mean	− 3.64e−04	− 1.48e−03	− 1.70e−05
$\sqrt{Var(R)}$	7.31e−03	6.28e−01	2.07e−03

R is defined by Eq. (11).