Table 5.

Regression and classification performance of MMD-DTA with varying similarity cutoffs

Similarity cutoff	Number of train sets	PCC	MSE	CI	Prec	Recall	BACC
1	313,414	0.844	0.946	0.838	0.890	0.789	0.836
0.6	289,107	0.809	0.881	0.819	0.820	0.876	0.822
0.59	259,290	0.718	1.298	0.800	0.836	0.824	0.815
0.58	226,521	0.763	1.083	0.793	0.834	0.793	0.801
0.57	183,555	0.678	1.531	0.749	0.826	0.665	0.748
0.56	148,252	0.587	1.940	0.707	0.814	0.566	0.705
0.55	113,519	0.595	1.756	0.708	0.776	0.628	0.705
0.54	81362	0.468	2.358	0.663	0.774	0.449	0.646
0.53	61,755	0.436	2.536	0.656	0.769	0.407	0.630
0.52	46,652	0.411	2.719	0.645	0.770	0.292	0.593
0.51	36,447	0.345	2.739	0.611	0.722	0.277	0.574
0.5	29,831	0.336	3.310	0.616	0.804	0.133	0.547

The number of test data points is fixed at 80,578