Table 3.

SVM modeling results based on different preprocesses.

Preprocessing Methods	Kernel Function	Best c	Best g	Rc	RMSEc	Rp	RMSEp	RPD
X	line	64.0000	\	0.91	0.3	0.90	0.3	2.4
SG		64.0000		0.91	0.3	0.90	0.3	2.4
MSC		2.8284		0.92	0.3	0.90	0.3	2.2
SNV		2.8284		0.92	0.3	0.89	0.3	2.2
1D		0.0442		0.90	0.3	0.91	0.3	2.2
1D + SG		0.0625		0.90	0.3	0.91	0.3	2.4
1D + MSC		0.0625		0.93	0.3	0.90	0.3	2.4
1D + SNV		0.0625		0.93	0.3	0.91	0.3	2.4
X	rbf	724.0773	0.0110	0.98	0.16	0.56	0.7	1.0
SG		724.0773	0.0110	0.97	0.16	0.56	0.7	1.0
MSC		90.5097	0.0010	0.86	0.4	0.82	0.4	1.8
SNV		512.0000	0.0010	0.93	0.3	0.84	0.4	1.8
1D		11.3137	0.0028	0.93	0.3	0.91	0.3	2.4
1D + SG		22.6274	0.0020	0.93	0.3	0.91	0.3	2.5
1D + MSC		2.8284	0.0156	0.98	0.14	0.84	0.4	1.9
1D + SNV		2.8284	0.0156	0.98	0.14	0.84	0.4	1.9