Table 3.
Summary of prediction performance of internal and external sets
| Dataset | Details | N | Acc (%) | Sen (%) | Spec (%) | MCC |
|---|---|---|---|---|---|---|
| Internal set (DPP4-TRN) | Full training | 1,122 | 96.43 | 98.30 | 94.38 | 0.929 |
| Ten-fold CV | 1,122 | 82.26 | 84.69 | 79.59 | 0.644 | |
| External set 1 (DPP4-TEST1) | External validation | 149 | 91.28 | – | – | – |
| External set 2 (DPP4-TEST2) | External validation | 160 | 95.63 | – | – | – |
| External set 3 (DPP4-TEST3) | External validation | 167 | 72.25 | – | – | – |
Note: N is the number of compounds.
Abbreviations: Acc, accuracy; CV, cross-validation; MCC, Matthews correlation coefficient; Sen, sensitivity; Spec, specificity.