Comparing the Potential of Near- and Mid-Infrared Spectroscopy in Determining the Freshness of Strawberry Powder from Freshly Available and Stored Strawberry

Table 1

Prediction result of strawberry storage days based on NIRⁿ spectra with full variables and optimal variables.

Model	Variable selection	Variable number	Calibration			Prediction			RPD^l	AB_RMSE^m
					RMSEC^h			RMSEP^k
PLSR^a	Full-variable	2074	0.959	0.920	2.790	0.953	0.908	2.933	3.302	0.142
LS_SVM^b	Full-variable	2074	1.000	1.000	0.001	0.973	0.945	2.264	4.302	2.263
PLSR	SPA^c	3	0.958	0.918	2.816	0.937	0.877	3.386	2.855	0.570
LS_SVM	SPA	3	0.977	0.954	2.113	0.970	0.940	2.359	4.123	0.246
PLSR	UVE^d	301	0.991	0.981	1.348	0.936	0.875	3.409	2.838	2.061
LS_SVM	UVE	301	0.996	0.991	0.913	0.984	0.968	1.729	5.680	0.816
PLSR	UVE_SPA	9	0.991	0.981	1.348	0.936	0.875	3.409	2.838	2.061
LS_SVM	UVE_SPA	9	0.996	0.992	0.872	0.985	0.968	1.723	5.699	0.852
PLSR	CARS^e	26	0.978	0.956	2.073	0.953	0.906	2.955	3.314	0.882
LS_SVM	CARS	26	0.993	0.986	1.176	0.973	0.944	2.289	4.322	1.113

^aPartial least squares regression, ^bleast squares support vector machines, ^csuccessive projections algorithm, ^duninformative variable elimination, ^ecompetitive adaptive reweighted sampling, ^fcorrelation coefficients of calibration, ^gdetermination coefficients of calibration, ^hroot-mean-square error of calibration, ⁱcorrelation coefficients of prediction, ^jdetermination coefficients of prediction, ^kroot-mean-square error of prediction, ^lresidual prediction deviation, ^mabsolute difference between RMSEC and RMSEP, and ⁿnear-infrared.