TABLE 2.

The classification error rate, specificity, and sensitivity of the partial least squares discriminant analysis (PLS-DA) models constructed for chocolate browning (CB) and friction browning (FB) of ‘Regal Seedless’ grape bunches.

Condition	Sample set	Class 0			Class 1
		CERa	Specb	Senc	CER	Spec	Sen
CB: W3and4d	Calibration	0.15	0.865	0.815	0.15	0.815	0.865
CB:W3and4	CVe	0.25	0.808	0.692	0.25	0.692	0.808
CB: W5and6f	Calibration	0.13	0.875	0.864	0.13	0.864	0.875
CB: W5and6	CV	0.22	0.722	0.818	0.22	0.818	0.722
FB: W3and4	Calibration	0.41	0.412	0.757	0.41	0.757	0.412
FB: W3and4	CV	0.46	0.353	0.714	0.26	0.714	0.353

^aClass error rate is defined as the proportion of instances misclassified over the whole set of instances.

^bSpecificity is defined as the ability of a test to correctly identify a sample without the defect.

^cSensitivity is defined as the ability of a test to correctly identify a sample with a defect.

^dWeeks 3 and 4.

^eCross-validation.

^fWeeks 5 and 6.