Table 4.
Detailed accuracy by class of the grapevine varietal classification using support vector machines and an external validation of 40 samples (n = 40).
| Class | True Positive Rate | False Positive rate | Precision | AUC |
|---|---|---|---|---|
| Cabernet Sauvignon (CS) | 1.000 | 0.000 | 1.000 | 1.000 |
| Caladoc (CL) | 1.000 | 0.000 | 1.000 | 1.000 |
| Carmenere (CR) | 1.000 | 0.000 | 1.000 | 1.000 |
| White Grenache (WG) | 1.000 | 0.028 | 0.800 | 0.993 |
| Pedro Ximenez (PX) | 1.000 | 0.000 | 1.000 | 1.000 |
| Pinot Noir (PN) | 0.750 | 0.028 | 0.750 | 0.972 |
| Tempranillo (TE) | 0.750 | 0.000 | 1.000 | 1.000 |
| Treixadura (TR) | 1.000 | 0.000 | 1.000 | 1.000 |
| Viognier (VO) | 0.750 | 0.000 | 1.000 | 1.000 |
| Viura (VU) | 1.000 | 0.028 | 0.800 | 1.000 |
| Weighted average | 0.925 | 0.008 | 0.935 | 0.997 |
AUC: area under the receiver operating characteristic (ROC) curve.