Table 4.
The performance of fingerprints based models developed using different machine learning techniques.
| Machine learning techniques (parameters) | Main dataset |
Validation dataset |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Sen | Spc | Acc | MCC | AUROC | Sen | Spc | Acc | MCC | AUROC | |
| SVM full feature (g = 0.001, c = 4, j = 1) | 95.91 | 87.25 | 91.62 | 0.84 | 0.97 | 93.16 | 86.56 | 89.89 | 0.80 | 0.97 |
| SVM after feature selection (g = 0.1, c = 6, j = 1) | 82.85 | 80.67 | 81.77 | 0.64 | 0.87 | 82.63 | 75.81 | 79.26 | 0.59 | 0.84 |
| Random Forest (Ntree = 100) | 92.88 | 90.07 | 91.48 | 0.83 | 0.98 | 92.63 | 89.25 | 90.96 | 0.82 | 0.97 |
| SMO (g = 0.001, c = 4) | 91.29 | 89.66 | 90.49 | 0.81 | 0.90 | 89.47 | 90.32 | 89.89 | 0.80 | 0.90 |
| J48 (c = 0.4, m = 1) | 90.50 | 88.99 | 89.75 | 0.80 | 0.88 | 88.95 | 86.02 | 87.50 | 0.75 | 0.85 |
| Naive Bayes (Default) | 84.30 | 64.56 | 74.52 | 0.50 | 0.74 | 78.42 | 65.05 | 71.81 | 0.44 | 0.72 |
Sen, Sensitivity; Spc, Specificity; Acc, Accuracy; MCC, Matthew’s Correlation Coefficient; AUROC, Area Under the Receiver Operating Characteristic curve.