Table 5.

The predictions of the best HCT116 QSAR model by the ten structural clusters for training and test sets. The best models are highlighted in bold.

Clusters	Training Set				Test Set
	#	R2	RMSE	MAE	#	R2	RMSE	MAE
I	650	0.702	0.721	0.479	129	0.613	0.673	0.455
II	569	0.766	0.627	0.446	135	0.781	0.619	0.461
III	438	0.792	0.648	0.459	105	0.697	0.737	0.516
IV	713	0.759	0.703	0.459	200	0.682	0.821	0.533
V	885	0.685	0.658	0.481	208	0.658	0.734	0.489
VI	405	0.649	0.646	0.460	119	0.637	0.616	0.462
VII	661	0.790	0.652	0.445	159	0.776	0.625	0.430
VIII	626	0.636	0.708	0.512	152	0.706	0.585	0.432
IX	513	0.846	0.599	0.412	125	0.794	0.720	0.487
X	415	0.746	0.628	0.440	132	0.767	0.659	0.448