Table 1.
Comparison of the proposed PSPred-ALE and existing methods on training and testing datasets.
| Model | 5-Fold cross validation |
Independent testing |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ACC (%) | AUC | SE (%) | SP (%) | MCC | ACC (%) | AUC | SE (%) | SE (%) | MCC | |
| DeepPSP | 80.21 | 0.876 | 76.65 | 83.78 | 0.606 | |||||
| MusiteDeep2017 | 80.17 | 0.880 | 78.87 | 81.46 | 0.604 | |||||
| MusiteDeep2020 | 80.95 | 0.887 | 82.95 | 78.96 | 0.620 | |||||
| DeepIPs | 80.45 | 0.887 | 79.70 | 81.19 | 0.610 | 80.63 | 0.894 | 79.61 | 83.50 | 0.632 |
| Ours (ψ = 33) | 83.58 | 0.911 | 83.89 | 83.26 | 0.671 | 83.06 | 0.907 | 82.58 | 83.54 | 0.661 |
| Ours (ψ = 29) | 84.03 | 0.913 | 84.22 | 83.84 | 0.681 | 83.14 | 0.907 | 83.15 | 83.13 | 0.663 |