Table 2.
Classification results and comparison
| Sensitivity (%) | Specificity (%) | Accuracy (%) | |
|---|---|---|---|
| Pearson's correlation | |||
| SVR | 63.9 | 95.5 | 80.6 |
| XGBoost | 86.8 | 89.7 | 88.4 |
| Linear PCA | |||
| SVR | 72.0 | 80.3 | 76.0 |
| XGBoost | 75.4 | 80.1 | 78.3 |
| Gaussian kernel PCA | |||
| SVR | 73.7 | 75.0 | 74.4 |
| XGBoost | 77.0 | 73.5 | 75.2 |
| Riemann kernel PCA | |||
| SVR | 86.6 | 85.1 | 86.6 |
| XGBoost | 92.6 | 90.7 | 91.8 |
These results are 10‐fold validated. Linear PCA and Riemann kernel PCA models were fitted with training data and then applied on validation data during 10‐fold validation. Bold values are the highest values in each column.