Table 3.

Cross-validation scores for the binary PLS-DA classification models built with different cut-off values.

	IC50-PLS-DA	CC50-PLS-DA
cut-off (mg/mL)	0.15	2.06
ONPC	7	3
EV	98%	78%
FNER	0.97	1.00
CVNER	0.85	0.97
ACC	0.87	0.97

ONPC: optimal number of PCs (see text). EV: explained variance. FNER: fitting-non-error-rate. CVNER: cross-validation-non-error-rate. ACC: accuracy.