Table 1. Results for CDK Dataset (Model Dataset 1) with Best Weighting (0.5–0–0.5) Combination as Obtained from the Retrospective Study.

model	division method	R2	Q2	MAE95%Train	QExt-F12	MAE95%Test	correct prediction in %	training set range	number of compounds outside ADa
1	sorted responseb	0.74	0.71	0.15	0.73	0.14	87.84	2.396	0
2		0.71	0.69	0.14	0.69	0.13	86.49		0
3		0.68	0.66	0.14	0.68	0.13	85.14		1 (122)
4		0.70	0.67	0.14	0.67	0.13	85.14		4 (118, 119, 122, 123)
5		0.63	0.61	0.17	0.65	0.17	77.02		0
6	Kennard–Stonec	0.78	0.75	0.14	0.63	0.13	85.29	2.342	0
7		0.77	0.75	0.14	0.59	0.14	79.41		1 (195)
8		0.74	0.71	0.14	0.54	0.13	89.71		0
9		0.72	0.70	0.14	0.56	0.12	83.82		0
10		0.71	0.68	0.14	0.54	0.13	86.76		0
11	modified-k-medoidsd	0.74	0.71	0.13	0.70	0.17	86.67	2.397	0
12		0.69	0.66	0.13	0.65	0.14	89.33		3 (30, 38, 39)
13		0.69	0.66	0.13	0.65	0.14	88		0
14		0.70	0.68	0.14	0.68	0.13	92		0
15		0.64	0.61	0.16	0.65	0.14	88		0

^aAD using standardization technique is used and compound ID mentioned under parenthesis.

^bN_Training = 154, N_Test = 74.

^cN_Training = 156, N_Test = 68.

^dN_Training = 149, N_Test = 75.