. 2020 Aug 7;9(8):1078. doi: 10.3390/foods9081078

Table 9.

Best results of each area with second derivative data and their preprocessing at a spectral sampling range of 729–975 nm using support vector machine regression (SVM-R).

Area	Pre-Processing	Calibration		Prediction (Validation)			RPD
Area	Pre-Processing	R²	RMSEC	R²	RMSEP	Bias	RPD
G	Autoscale	0.706	1.8103	0.683	1.8424	−0.146	1.7694
M	Autoscale	0.823	1.7344	0.803	1.8421	−0.064418	2.2691
S	Autoscale	0.802	1.7238	0.790	1.6644	0.32509	2.1629
W	Autoscale	0.792	1.0415	0.754	1.1596	−0.18288	2.0006
Y	Autoscale	0.755	0.91677	0.694	0.9714	−0.008294	2.3883