Table 2.

PLS regression results using RCGA-PLS variable selection method

Parameters	LVs	RMSEC	RC2	RMSECV	RP2	RMSEP
20	2	2.8092	0.6194	2.9256	0.706	2.5248
	3	2.6389	0.6641	2.7517	0.7492	2.332
	4	2.1943	0.7678	2.4655	0.7293	2.4229
	5	3.1093	0.5337	3.779	0.4445	3.4706
	6	2.3253	0.7392	2.7954	0.6647	2.6965
	7	2.0814	0.791	2.6516	0.7548	2.3057
	8	2.5253	0.6924	3.1198	0.4727	3.3814
30	2	0.4608	0.9898	0.4763	0.9915	0.4301
	3	0.3431	0.9943	0.3687	0.993	0.3907
	4	0.1983	0.9981	0.2172	0.9953	0.3205
	5	0.1871	0.9983	0.2135	0.9969	0.2607
	6	0.1867	0.9983	0.2195	0.9962	0.2865
	7	0.1916	0.9982	0.2425	0.9943	0.3526
	8	0.1737	0.9985	0.2182	0.9965	0.2773
40	2	0.4685	0.9894	0.4811	0.9892	0.4842
	3	0.2874	0.996	0.306	0.9958	0.3035
	4	0.4495	0.9903	0.5521	0.9866	0.538
	5	0.3887	0.9927	0.4505	0.9817	0.6306
	6	0.2804	0.9962	0.3671	0.9923	0.4084
	7	0.5432	0.9858	0.713	0.9807	0.6473
	8	0.522	0.9869	0.6767	0.9665	0.8529
50	2	0.2408	0.9972	0.2502	0.9972	0.2455
	3	0.21	0.9979	0.2221	0.9974	0.2365
	4	0.1777	0.9985	0.2036	0.9973	0.2411
	5	0.1776	0.9985	0.2047	0.9961	0.2912
	6	0.1611	0.9987	0.2026	0.9961	0.2921
	7	0.1507	0.9989	0.1919	0.9965	0.2756
	8	0.1225	0.9993	0.1653	0.996	0.2961
60	2	0.2191	0.9977	0.2286	0.9977	0.2233
	3	0.193	0.9982	0.2069	0.9978	0.2171
	4	0.1759	0.9985	0.2023	0.9977	0.2241
	5	0.1642	0.9987	0.1993	0.9968	0.263
	6	0.1413	0.999	0.172	0.9968	0.2627
	7	0.138	0.9991	0.1773	0.9965	0.2737
	8	0.1124	0.9994	0.1555	0.9965	0.2764

^aThe data size for training and testing were 100 and 60, respectively, partitioning of data is using SPXY, cross-validation was based on kfold = 10, and the preprocessing of the data was SNV + SG1