Table 4.

Prediction of oil content and fatty acid contents in mustard seeds from iPLS model using Vis–NIR and SWIR HSI spectra.

	Spectral pre-processing	NW	LV	Calibration		Cross validation		Testing
				$R_C^2$	RMSEC	$R_{\mathit{CV}}^2$	RMSECV	$R_P^2$	RMSEP
Vis–NIR
Oil	SNV	14	3	0.877	1.139	0.877	1.140	0.748	1.585
Fatty acids
C18:1	SNV	12	5	0.875	0.168	0.874	0.168	0.763	0.225
C18:2	SNV	20	9	0.659	0.281	0.653	0.283	0.645	0.275
C20:1	Raw	50	8	0.850	0.060	0.847	0.061	0.804	0.069
C18:3	SG smoothing	74	9	0.850	0.298	0.848	0.301	0.816	0.330
C22:1	Raw	48	8	0.771	0.619	0.767	0.625	0.784	0.598
Saturated	SNV	12	3	0.834	0.172	0.834	0.172	0.777	0.184
Unsaturated	SNV	12	4	0.836	0.171	0.835	0.172	0.777	0.184
MUFA	SNV	32	9	0.613	0.528	0.606	0.533	0.595	0.536
PUFA	Raw	64	12	0.725	0.424	0.716	0.431	0.691	0.446
SWIR
Oil	SNV + SG-D2	72	6	0.884	1.060	0.880	1.076	0.885	1.058
Fatty acids
C18:1	SNV + SG-D2	38	12	0.882	0.158	0.874	0.163	0.856	0.176
C18:2	SNV	12	4	0.780	0.208	0.777	0.209	0.715	0.247
C20:1	SNV	82	10	0.874	0.057	0.864	0.059	0.836	0.063
C18:3	SNV	84	9	0.871	0.283	0.863	0.291	0.866	0.288
C22:1	SNV + SG smoothing	74	12	0.861	0.501	0.856	0.510	0.861	0.501
Saturated	SNV	12	4	0.925	0.102	0.925	0.103	0.828	0.162
Unsaturated	Raw	26	7	0.821	0.157	0.820	0.158	0.802	0.174
MUFA	SNV	56	10	0.848	0.337	0.840	0.345	0.812	0.354
PUFA	Raw	32	8	0.793	0.381	0.791	0.383	0.753	0.399

LV latent variables, NW number of wavelengths, R² coefficient of determination, RMSEC root mean square error for calibration, RMSECV root mean square error for cross validation, RMSEP root mean square error for testing.