Table 3 Comparison results of the EviDTI and baselines on the Davis dataset
From: Evidential deep learning-based drug-target interaction prediction
Davis | Accuracy (std) | Recall (std) | Precision (std) | MCC (std) | F1 score (std) | AUC (std) | AUPR (std) |
|---|---|---|---|---|---|---|---|
RF | 84.48 (0.60) | 65.01 (1.47) | 80.70 (1.87) | 64.12 (2.11) | 72.01 (1.49) | 89.64 (0.68) | 83.04 (1.68) |
NB | 67.53 (0.92) | 61.70 (1.82) | 47.78 (1.95) | 45.21 (1.68) | 53.85 (1.72) | 70.11 (1.39) | 58.43 (2.09) |
SVM | 82.13 (0.59) | 59.70 (1.57) | 76.96 (2.19) | 60.23 (1.45) | 67.24 (1.61) | 87.85 (0.78) | 80.16 (1.65) |
DeepConv-DTI | 86.41 (0.27) | 72.14 (0.33) | 77.85 (0.44) | 64.53 (0.35) | 74.89 (0.64) | 91.83 (0.20) | 83.87 (0.62) |
GraphDTA | 81.72 (0.19) | 53.01 (1.76) | 74.33 (1.45) | 62.42 (0.67) | 71.83 (0.82) | 85.92 (0.43) | 74.31 (0.74) |
MolTrans | 83.61 (0.18) | 76.31 (0.34) | 68.87 (0.24) | 62.91 (0.13) | 72.40 (0.17) | 89.84 (0.22) | 80.26 (0.66) |
HyperAttention | 86.61 (0.14) | 78.01 (0.23) | 75.41 (0.16) | 67.54 (0.28) | 74.21 (0.22) | 92.01 (0.14) | 83.92 (0.13) |
TransformerCPI | 82.22 (0.16) | 68.82 (0.38) | 68.81 (0.37) | 64.21 (0.24) | 72.41 (0.25) | 87.72 (0.16) | 76.71 (0.18) |
GraphormerDTI | 84.81 (0.84) | 83.78 (2.52) | 68.43 (2.18) | 65.22 (1.84) | 75.28 (1.43) | 91.24 (0.74) | 84.31 (1.48) |
AIGO-DTI | 86.19 (0.79) | 72.55 (2.94) | 78.01 (2.11) | 65.72 (1.96) | 75.12 (1.52) | 92.01 (0.97) | 84.81 (1.10) |
DLM-DTI | 85.66 (0.47) | 74.08 (2.50) | 75.26 (1.92) | 64.68 (1.27) | 74.62 (1.03) | 91.30 (0.35) | 82.48 (0.86) |
EviDTI | 87.40 (1.08) | 72.93 (1.93) | 81.17 (2.94) | 68.40 (2.78) | 76.81 (2.03) | 92.10 (0.78) | 85.07 (2.12) |
