. 2021 Jun 21;61(7):3255–3272. doi: 10.1021/acs.jcim.1c00451

Table 4. Average Performance of the CP Models Generated from a Selected Set of Features^a.

endpoint	descriptor	validity	STD validity	efficiency	STD efficiency	F1 score	STD F1 score	MCC	STD MCC
MNT	CHEM	0.77	0.02	0.76	0.05	0.61	0.02	0.28	0.05
	BIO	0.82	0.03	0.81	0.05	0.70	0.03	0.46	0.06
	CHEMBIO	0.81	0.03	0.85	0.03	0.70	0.03	0.44	0.07
DILI	CHEM	0.78	0.05	0.91	0.04	0.74	0.05	0.49	0.09
	BIO	0.81	0.04	0.83	0.07	0.76	0.04	0.53	0.07
	CHEMBIO	0.81	0.03	0.88	0.04	0.77	0.03	0.55	0.06
DICC	CHEM	0.79	0.02	0.84	0.02	0.72	0.03	0.46	0.05
	BIO	0.79	0.02	0.96	0.02	0.81	0.01	0.63	0.02
	CHEMBIO	0.79	0.02	0.94	0.01	0.82	0.01	0.65	0.03

Mean and standard deviation (STD) calculated over a 5-fold CV. The highest mean per metric and endpoint is highlighted (bold).

Table 4. Average Performance of the CP Models Generated from a Selected Set of Featuresa.