Table 11 Comparison of results between KNN classifier with SVM
From: Prediction of acid radical ion binding residues by K-nearest neighbors classifier
Ligand | Method | L | Optimal k value | Sn (%) | Sp (%) | Acc (%) | FPR (%) | MCC |
|---|---|---|---|---|---|---|---|---|
Zn2+ | OUR’S | 7 | 103 | 94.3 | 83.8 | 89.1 | 16.2 | 0.786 |
SVM | – | 99.8 | 99.5 | 99.7 | – | 0.993 | ||
Fe2+ | OUR’S | 9 | 41 | 92.1 | 80.4 | 86.3 | 19.6 | 0.730 |
SVM | – | 91.9 | 90.7 | 91.3 | – | 0.826 | ||
Fe3+ | OUR’S | 9 | 15 | 84.6 | 84.9 | 84.7 | 15.1 | 0.694 |
SVM | – | 86.9 | 88.7 | 87.8 | – | 0.756 | ||
Cu2+ | OUR’S | 13 | 49 | 92.4 | 86.6 | 89.5 | 13.4 | 0.791 |
SVM | – | 95.5 | 97.1 | 96.3 | – | 0.926 | ||
Mn2+ | OUR’S | 7 | 23 | 79.1 | 80.9 | 80.0 | 19.1 | 0.600 |
SVM | – | 82.1 | 84.4 | 83.2 | – | 0.664 | ||
Co2+ | OUR’S | 11 | 99 | 77.6 | 83.1 | 80.3 | 16.9 | 0.608 |
SVM | – | 80.8 | 85.1 | 83.0 | – | 0.660 |