Table 5.

Parameters and prediction differences of water models.

Parameter	Water Model 1		Water Model 2		Water Model 3		Water Model 4
	Training set	Test set	Training set	Test set	Training set	Test set	Training set	Test set
Number of samples	47	19	46	16	30	10	32	11
Content range (%, mg/mg)	1.39–9.29	1.50–7.17	0.41–8.98	1.29–7.67	1.42–8.73	1.89–7.08	0.50–2.60	1.30–2.28
Wavenumber range (cm−1)	9403.5–7498.1		10502.8–9951.2, 7752.7–5546.4		10502.8–8848.1, 8304.2–7197.3		8851.9–7748.8, 6101.9–5546.4
Pre-processing method	1st derivative + vector normalization		1st derivative + vector normalization		1st derivative + vector normalization		1st derivative + vector normalization
Rank	3		5		6		2
R2 (%)	99.33		99.67		99.23		93.59
RMSEP (%)	0.174		0.117		0.13		0.0737
Samples whose prediction difference is >5%	0% (0/377)		0% (0/377)		0% (0/78)		0% (0/96)
Samples whose prediction difference is <1%	87.0% (328/377)		80.9% (305/377)		47.4% (37/78)		96.9% (93/96)
Average prediction difference	0.4%		0.5%		1.0%		0.4%