Table 1.

Values of the AUROC metric for the best eight individual models and the best 8-model ensemble.

Model	Training set	Test set	DUD-E database
8-MODEL ENSEMBLE (MIN)	0.851 (±0.0281)	0.885 (±0.0367)	0.976*** (±0.0085)
8-MODEL ENSEMBLE (RANKING)	0.886 (±0.0239)	0.878 (±0.0375)	0.976*** (±0.0082)
8-MODEL ENSEMBLE (AVERAGE)	0.891 (±0.0234)	0.887 (±0.0357)	0.970*** (±0.0096)
8-MODEL ENSEMBLE (VOTING)	0.833 (±0.0283)	0.810 (±0.0516)	0.959* (±0.0173)
348	0.882 (±0.0250)	0.885 (±0.0360)	0.934 (±0.0123)
706	0.809* (±0.0319)	0.736** (±0.0565)	0.924 (±0.0205)
981	0.843 (±0.0298)	0.837 (±0.0467)	0.922 (±0.0254)
557	0.778** (±0.0343)	0.818 (±0.0482)	0.919 (±0.0203)
123	0.850 (±0.0285)	0.882 (±0.0382)	0.918 (±0.0185)
693	0.860 (±0.0280)	0.828* (±0.0459)	0.913* (±0.0171)
560	0.775** (±0.0348)	0.779* (±0.0525)	0.911* (±0.0185)
746	0.844 (±0.0288)	0.820 (±0.0456)	0.910 (±0.0195)

AUROCs statistically different from the correspondent column for the best individual model (model 348).

^*p < 0.05,

^**p < 0.01, and

^***p < 0.001.

The highest AUC for an individual model is indicated in bold.