Table 3 Classification performance of the different models
Models | Training dataset (n = 96) | Test dataset (n = 24) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Accuracy (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUC | Accuracy (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUC | |
Intra-voxel | ||||||||||||
DT | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 54.17 | 64.29 | 40.00 | 60.00 | 44.44 | 0.52 |
RF | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 75.00 | 64.29 | 90.00 | 90.00 | 64.29 | 0.69 |
XGBoost | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 75.00 | 71.43 | 80.00 | 83.33 | 66.67 | 0.72 |
Inter-voxel | ||||||||||||
DT | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 54.17 | 57.14 | 50.00 | 61.54 | 45.45 | 0.54 |
RF | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 66.67 | 71.43 | 60.00 | 71.43 | 60.00 | 0.58 |
XGBOOST | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 62.50 | 100.00 | 10.00 | 60.87 | 100.00 | 0.54 |
Combined intra- and intervoxel | ||||||||||||
DT | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 66.67 | 57.14 | 80.00 | 80.00 | 57.14 | 0.69 |
RF | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 75.00 | 78.57 | 70.00 | 78.57 | 70.00 | 0.71 |
XGBoost | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 1.00 | 91.67 | 92.86 | 90.00 | 92.86 | 90.00 | 0.94 |