Table 3 Performance of different coding methods
From: SDNN-PPI: self-attention with deep neural network effect on protein-protein interaction prediction
Encoding methods | Length | ACC(%) | Spec(%) | Sens(%) | Prec(%) | MCC(%) | AUC(%) |
|---|---|---|---|---|---|---|---|
ACC+CT+LD+AC | 1203 | 92.31 ± 0.66 | 94.37 ± 0.25 | 90.26 ± 1.28 | 94.13 ± 0.25 | 84.70 ± 1.27 | 97.00 |
ACC+CT+LD | 993 | 92.00 ± 0.66 | 93.31 ± 1.04 | 90.69 ± 0.57 | 93.14 ± 1.01 | 84.03 ± 1.33 | 97.03 |
ACC+CT+AC | 573 | 95.19 ± 0.68 | 97.05 ± 0.65 | 93.33 ± 1.20 | 96.94 ± 0.65 | 90.45 ± 1.34 | 98.60 |
ACC+LD+AC | 860 | 91.41 ± 0.52 | 93.06 ± 0.87 | 89.76 ± 0.64 | 92.83 ± 0.86 | 82.87 ± 1.06 | 96.58 |
CT+LD+AC | 1183 | 89.50 ± 0.68 | 90.97 ± 0.58 | 88.02 ± 1.47 | 90.70 ± 0.50 | 79.04 ± 1.33 | 95.62 |
ACC+CT | 363 | 89.79 ± 0.65 | 90.79 ± 0.81 | 88.79 ± 0.87 | 90.61 ± 0.76 | 79.61 ± 1.29 | 95.78 |
ACC+LD | 650 | 88.93 ± 0.43 | 89.67 ± 1.37 | 88.20 ± 1.53 | 89.54 ± 1.14 | 77.91 ± 0.88 | 95.05 |
ACC+AC | 230 | 85.74 ± 1.80 | 87.20 ± 1.74 | 84.29 ± 2.66 | 86.82 ± 1.74 | 71.54 ± 3.60 | 92.59 |
CT+LD | 973 | 90.47 ± 0.52 | 92.76 ± 0.57 | 88.18 ± 0.93 | 92.42 ± 0.55 | 81.03 ± 1.02 | 95.97 |
CT+AC | 553 | 89.06 ± 0.55 | 90.70 ± 0.87 | 87.43 ± 1.34 | 90.40 ± 0.73 | 78.19 ± 1.07 | 94.74 |
LD+AC | 840 | 91.44 ± 0.44 | 92.72 ± 0.72 | 90.17 ± 0.56 | 92.54 ± 0.69 | 82.92 ± 0.89 | 96.60 |