Research Article
Ensemble Learning Prediction of Drug-Target Interactions Using GIST Descriptor Extracted from PSSM-Based Evolutionary Information
Table 5
5-fold cross-validation results were generated through the proposed method on the nuclear receptor dataset.
| Testing set | Accuracy (%) | Precision (%) | Sensitivity (%) | MCC (%) | AUC | AUPR |
| 1 | 75.00 | 76.92 | 62.50 | 60.78 | 0.7688 | 0.6444 | 2 | 80.56 | 78.95 | 83.33 | 68.62 | 0.8148 | 0.7007 | 3 | 69.44 | 81.25 | 61.90 | 56.69 | 0.8254 | 0.8484 | 4 | 72.22 | 60.00 | 85.71 | 58.61 | 0.8442 | 0.7507 | 5 | 72.22 | 72.00 | 85.71 | 56.06 | 0.7524 | 0.7053 | Average | 73.89 ± 4.21 | 73.82 ± 8.45 | 75.83 ± 12.48 | 60.15 ± 5.08 | 0.8011 ± 0.0389 | 0.7299 ± 0.0762 |
|
|