. 2021 Jan 12;7(2):406–420. doi: 10.1021/acsinfecdis.0c00754

Table 3. Five-Fold Cross-Validation Results of Bayesian Machine Learning Models Used to Predict 56 Compounds for Screening against S. mansoni In Vitro^a.

time point (h)	life stage	no. active	no. total	ROC	F1-score	CK	MCC	domain
≤24	adults	45	493	0.845	0.426	0.339	0.397	0.308
≤24	somules	227	3114	0.806	0.260	0.157	0.246	0.383
48	adults	35	450	0.832	0.310	0.210	0.307	0.307
48	somules	204	3101	0.818	0.276	0.187	0.276	0.382
72	adults	34	447	0.841	0.494	0.441	0.457	0.305
72	somules	175	2979	0.843	0.276	0.198	0.290	0.377
>72	adults	85	509	0.815	0.533	0.410	0.432	0.308
>72	somules	358	3151	0.796	0.366	0.232	0.302	0.384

Terms: ROC = receiver operator characteristic; CK = Cohen’s kappa; MCC = Matthews correlation coefficient.

Table 3. Five-Fold Cross-Validation Results of Bayesian Machine Learning Models Used to Predict 56 Compounds for Screening against S. mansoni In Vitroa.