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

Table 2

Prediction result of strawberry storage days based on MIRⁿ 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	1762	0.988	0.977	1.505	0.981	0.962	1.876	5.155	0.371
LS_SVM^b	Full-variable	1762	1.000	1.000	0.027	0.994	0.985	1.183	8.436	1.156
PLSR	SPA^c	7	0.984	0.969	1.731	0.982	0.964	1.823	5.296	0.092
LS_SVM	SPA	7	0.999	0.999	0.359	0.996	0.992	0.864	11.198	0.505
PLSR	UVE^d	230	0.994	0.988	1.094	0.984	0.968	1.718	5.621	0.625
LS_SVM	UVE	230	0.996	0.993	0.833	0.989	0.978	1.417	6.817	0.583
PLSR	UVE_SPA	12	0.994	0.988	1.094	0.984	0.968	1.718	5.621	0.625
LS_SVM	UVE_SPA	12	0.997	0.993	0.812	0.989	0.979	1.413	6.835	0.601
PLSR	CARS^e	32	0.979	0.957	2.034	0.954	0.907	2.951	3.330	0.917
LS_SVM	CARS	32	0.991	0.983	1.288	0.976	0.948	2.190	4.600	0.902

^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 ⁿmid-infrared.