Table 7.

Golbraikh and Tropsha statistical criteria to validate the designed QSAR model.

Parameter	Equation	Model Score	Threshold	Comment
$R^2$	$R^2=1-\frac{{\displaystyle \sum } (Yobs-Y cal ){ }^2}{{\displaystyle \sum } \left(Yobs-\overline{Yobs}\right){ }^2}$	0.69	>0.6	Accepted
$R^{2}adj$	$R^{2}adj=\frac{\left(N-1\right)R{ }^2-p}{N-p-1}$	0.62	>0.6	Accepted
$R^{2}test$	$R^{2}test=1-\frac{{\displaystyle \sum } (Ycal\left(test\right)-Y obs\left(test\right) ){ }^2}{{\displaystyle \sum } \left(Yobs\left(test\right)-\overline{Yobs}\left(train\right)\right){ }^2}$	0.63	>0.6	Accepted
$Q^{2}cv$	$Q^{2}cv=1-\frac{{\displaystyle \sum } (Ycal-Y obs ){ }^2}{{\displaystyle \sum } \left(Yobs-\overline{Yobs}\right){ }^2}$	0.57	>0.5	Accepted
R2 rand	Average of the 100 R2 rand (i)	0.17	<R2	Accepted
$Q^{2}cv$ ‘LOO’ rand	Average of the 100 $Q^{2}cv$ ‘LOO‘ rand (i)	−0.34	<Q2cv	Accepted
cR2p	cR2p = R*$\sqrt[2]{(R^2-\left( Average R_{rand} ){ }^2\right)}$	0.60	>0.5	Accepted

Yobs and Ycalc: refer to the observed and calculated/predicted response values. $\overline{Yobs}$ and $\overline{Ycal}$ refer to the mean of the observed and calculated/predicted response values. N and p refer to the number of data points (compounds) and descriptors.