Table 6. Results of PLSR models for TNCs-HSI analysis based on full spectra (F-PLSR) and the selected important wavelengths (RF-PLSR).

Samples	Models	N.[1]	LVs	Calibration		Cross-Validation		Prediction
				RC	RMSEC (%)	RCV	RMSECV (%)	RP	RMSEP (%)
Leaf	F-PLSR[5]	460	17	0.974	0.156	0.924	0.269	0.934	0.223
	RF-PLSR[6]	10	3	0.817	0.401	0.798	0.420	0.828	0.356
Stem	F-PLSR	460	16	0.987	0.063	0.917	0.168	0.930	0.084
	RF-PLSR	9	5	0.820	0.229	0.773	0.254	0.724	0.157
Root	F-PLSR	460	11	0.992	0.020	0.931	0.064	0.915	0.045
	RF-PLSR	7	7	0.773	0.097	0.592	0.130	0.797	0.068
Whole-plant	F-PLSR	460	6	0.904	0.388	0.893	0.409	0.908	0.351
	RF-PLSR	8	4	0.878	0.451	0.874	0.468	0.876	0.426

Note:

^[1]N.: Number of wavelengths used for analysis;

^[5]F-PLSR models meant the PLSR models established by using full spectra;

^[6]RF-PLSR models represented the PLSR models built by important wavelengths selected by RF algorithm. LVs, R_C, RMSEC, R_CV, RMSECV, R_P, and RMSEP could be found in the text.