Table 2 Prediction performance for carbonization characteristics of PLSR models in NIR spectra with 1250–2300 nm region and comparison with OLSR models.

Output variable	Spectrum pretreatment	PLSR					OLSR
		PLS factors	Calibration		Prediction		Calibration		Prediction
		PLS factors	R²_c	RMSEC	R²_p	RMSEP	R²_c	RMSEC	R²_p	RMSEP
C (wt%)	Original	8	0.987	0.1840	0.973	0.2590	0.999	0.0003	0.953	0.3440
C (wt%)	2nd derivative	7	0.991	0.1486	0.976	0.2461	0.999	0.0003	0.940	0.3885
O/C	Original	6	0.974	0.0048	0.963	0.0057	0.999	0.0003	0.907	0.0091
O/C	2nd derivative	5	0.981	0.0041	0.964	0.0056	0.999	0.0003	0.874	0.0106
H/C	Original	7	0.990	0.0003	0.984	0.0004	0.992	0.0003	0.966	0.0006
H/C	2nd derivative	6	0.994	0.0003	0.984	0.0004	0.992	0.0003	0.912	0.0010

PLSR partial least squares regression, OLSR ordinary least squares regression, R²_c coefficient of determination for calibration, R²_p coefficient of determination for prediction, RMSEC root mean square error of calibration, RMSEP root mean square error of prediction.

Quick links

Search