Table 1.

Comparisons among different quantitative structure-activity relationships (QSAR) models on ferric thiocyanate (FTC) dataset ^a.

Descriptors	Before Outlier Elimination			After Outlier Elimination
	Q2	R2	optPC	Q2	R2	optPC	Outlier
HSEHPCSV	0.3861	0.5781	4	0.6170	0.7338	20	183, 182, 181, 134
ST-scale	0.4268	0.5733	12	0.5993	0.6844	13	183, 182, 181, 134
HESH	0.4091	0.5366	2	0.5968	0.7047	10	183, 181, 182, 134, 129
VSW	0.4901	0.5771	3	0.5925	0.6768	5	181, 183, 182, 134, 151
G-scale	0.4516	0.5527	6	0.5843	0.6574	9	181, 183, 182, 134, 118
FASGAI	0.4814	0.5457	5	0.5544	0.6130	6	129, 181, 128
DPPS	0.4740	0.5637	7	0.5379	0.6278	8	181, 182, 183, 134
E-scale	0.4956	0.5451	4	0.5144	0.5582	4	181, 182, 183, 112
5Z-scale	0.3903	0.4626	12	0.3974	0.4653	9	181, 182, 183, 172
VHSE	0.4265	0.5432	12	0.3974	0.514	8	181, 182, 183, 172
T-scale	0.3280	0.4215	9	0.3728	0.4362	9	181, 182, 183
V-scale	0.3371	0.3785	5	0.3070	0.3458	6	181, 183, 182
Z-scale	0.2814	0.3398	4	0.2678	0.3415	4	181
ISA-ECI	0.1493	0.1916	6	0.1572	0.1836	6	183, 182, 181
MS-WHTM1	0.0736	0.1488	3	0.1036	0.1678	3	181, 183, 182
MS-WHTM2	0.0775	0.1445	3	0.0882	0.1617	3	181, 182, 183
Integrated descriptors	0.4811	0.5843	3	0.6818	0.7964	8	181, 183, 182, 134, 151, 153, 188
BOSS				0.7471 ± 0.0032	0.7931 ± 0.0062	9.72 ± 3.2199

^a R² is the coefficient of determination; Q² is the cross-validated R²; optPC is optimal principal components for PLS regression model; the results of BOSS are shown in the form of mean value ± standard deviation in 100 runs; the top ranked Q² scores were marked in bold.