Table 3.

Statistical evaluations of predictions for the Extra Trees, Random Forest and Bagging predictors and for the Vox Machinarum consensus classifier over our sourced literature logS values (see Table A2 for references) for 26 compounds from the 2019 Solubility Challenge loose test set of 32 molecules. The Vox Machinarum predictions reported here were the median of the other three classifiers’ predictions for each compound. The standard deviation of the 26 compounds’ logS values is 2.006.

Method	RMSE	RMSE/SD	AAE	R 2	Err < 0.5	Err < 1.0
Extra Trees	1.716	0.856	1.226	0.349	7 (27%)	12 (46%)
Random Forest	1.619	0.807	1.160	0.427	7 (27%)	15 (58%)
Bagging	1.558	0.777	1.119	0.482	7 (27%)	16 (62%)
Vox Machinarum	1.617	0.806	1.158	0.429	7 (27%)	15 (58%)