DRDB: A Machine Learning Platform to Predict Chemical-Protein Interactions towards Diabetic Retinopathy

Table 3

Performance summary of the multivoting ensemble method on integrated test set.

Cutoff	AUC		SE	SP	MCC
1	0.724	0.724	0.957	0.491	0.507
2	0.766	0.766	0.936	0.596	0.566
3	0.789	0.789	0.916	0.663	0.597
4	0.802	0.802	0.903	0.701	0.617
5	0.810	0.810	0.888	0.731	0.627
6	0.815	0.815	0.879	0.751	0.635
7	0.817	0.817	0.863	0.770	0.636
8	0.821	0.821	0.843	0.799	0.643
9	0.824	0.824	0.820	0.828	0.648
10	0.819	0.819	0.795	0.844	0.64
11	0.816	0.816	0.771	0.86	0.634
12	0.814	0.814	0.749	0.879	0.633
13	0.803	0.803	0.709	0.897	0.617
14	0.792	0.792	0.665	0.919	0.603
15	0.765	0.766	0.591	0.940	0.566
16	0.686	0.686	0.409	0.962	0.446